It's a tale as old as time that developers, particularly junior developers, are convinced they could "slap together something in one weekend" that would replace expensive SAAS software and "just do the parts of it we actually use". Unfortunately, the same arguments against those devs regular-coding a bespoke replacement apply to them vibe-coding a bespoke replacement: management simply doesn't want to be responsible for it. I didn't understand it before I was in management either, but now that I'm in management I 100% get it.
OTOH, I was hired by an enterprise that was many months into a giant backend rewrite. After wrapping my head around the many plans, I realized they were rewriting Django, badly. One weekend I prototyped the whole thing… in Django. It worked. It met the specs. It was a CRUD app with a REST API.
I came in to work Monday morning, showed it off, and inadvertently triggered a firestorm. Later my boss told me not to do that again because it caused havoc with schedules and such.
So I quit and found a better job. Sometimes the new guy can make a better version themselves over the weekend, not because they’re a supergenius, but because they’re not hampered by 47 teams all trying to get their stamp on the project.
(In before “prime example of overconfidence!”: feel free to doubt. It was a CRUD app with a handful of models on a PostgreSQL backend. They were writing a new Python web framework to serve it, complete with their own ORM and forms library and validation library. Not because the existing ones wouldn’t work, mind you, but more out of not realizing that all these problems were already sufficiently solved for their requirements.)
I can relate to this so much. When I was a newly joined Google consultant at a partner firm, we went to their office - some 13 different types of cuisines, different types of game rooms, lounges and what not. A luxury star hotel experience. We were waiting for our meeting on behalf of this one particularly large media client who was bleeding money on Wordpress.
3 engineers arrived - fashionably late. We explained them the situation and all we wanted from them was some GCP offering that would cure our woes and one that would cut our bills. The senior consultant - and presumably the only tech guy (rest seemed to be salesy) wasted our time like a used car salesman - he didn't even understand Google's own product portfolio and recommended us to use something like Spanner - which was totally not the solution to the problem, not to mention, expensive.
My boss and I left the meeting pissed off and he told me - "Neya, you probably know more about the product portfolio than these guys. Let's leave". That weekend, I went with my tried and trusted favorite Db - PostgreSQL - CloudSQL with a custom Elixir middleware based an old CMS I wrote a decade ago. After some trial and error, the solution worked flawlessly (and still does till date on auto-pilot). My client still has the lowest cost in the region - 1/3rd the cost of their competitors...7 years later. Back then, there was no vibe-coding, no AI, no auto-completion. Just pure thinking and experimentation.
All this just to say I agree that the new guy sometimes can make the best solutions to a problem and not always screw up. I always listen to new hires these days (now I'm a fractional / CTO) because you never know who could pull off that 1/3rd cost cutting framework move.
> Later my boss told me not to do that again because it caused havoc with schedules and such.
Did you talk to anyone about your plans before you brought in the demo or let them know they were solved problems? Often these sorts of reactions come down to your boss not wanting their team to lose their jobs because of the perception that it can all be handled by one person who's happy to work weekends.
Oh I have seen this story and was the one who caused this story when I was younger.
In a lot of cases the "new guy" thinks its an easy software and does it on his free time and thinks he did a great job.
In reality the specs are never 100% done correctly. The "new guy" misses some edge cases everybody but him knew because its just company knowledge. A lot of info in the specs was missing since they are not complete and so on.
This over the weekend never works in the long run. The ORM worked for all the happy path and written down cases but then you have cases were the ORM just is not good enough or fast. So you start to add strange code to work around the ORM. The same for the web framework or the validation lib.
To me the author of this comment sounds like the typical "Freelancer" coming in into a company knowing everything better then all the people and then leaving after a few months and now everybody else has to deal with his code.
> Sometimes the new guy can make a better version themselves over the weekend, not because they’re a supergenius, but because they’re not hampered by 47 teams all trying to get their stamp on the project.
In 99.9% of cases what seems to be "the better" version is better only for the "new guy" or rather his ego.
Those 47 teams hampering doesn't necessarily mean a bad thing, and more often than not actually well justified "stamps".
You only understand those things when you turn from the "supergenius" into an owner who have to take care not only of numbers on screen, but also security, interfacing, management and so on.
Take this idea and bring your own validation library and forms and UI components to the next job, and you've described what I do. And then you have real lock-in.
But I don't think Claude Code is going to prevent an org that thinks they can prompt their way to a replacement for all their SaaS from having internal political bickering that makes them end up with a extra-shitty mega-compromise to try to make all the internal stakeholders happy.
If you've got no vision and no taste, you need to find a vendor who will protect you from screwing up your internal processes and tools.
Internal tools teams have rarely cared much about UX or the day-to-day experience of their poor users. The quick-and-dirty internal-prompt-based one is likely to similarly be unimaginative and unintuitive.
Isn't this agreeing with the parent? If Django were the B2B SaaS product, you didn't vibe-code Django, you just used Django. You aren't responsible for maintaining Django itself.
I once worked on a project that was a new sign-up process for a large retailer (~90k employees at the time, four figures worth of outlets, quite a few billions in turnover).
There were about 60 people across, I can't even remember, maybe 10 teams that I knew about. One of the managers there assured me that the true participant figure was closer to 160. I was agog.
This project took months, and it had been going for months before I joined. And, as you can imagine, with that many people involved over that sort of timescale, there was turnover: people, sometimes key people, would leave and be replaced - sometimes by others who'd already worked on the project, but sometimes by new people - with the expected disruption that causes.
It was two web pages, plus some back end plumbing across multiple systems.
But, in the grand scheme of things, nothing that complex. I reckon it was a handful of thousands lines of code total across the full stack. I was mostly on the database side and, IIRC, I wrote a few hundred lines of SQL in a couple of stored procedures (wouldn't have been my preferred solution building from scracth but, you know, we weren't, and we had to integrate with a legacy systems that had to keep working as expected).
Overall it took 8 or 9 months from kick off to shipping and I was involved for perhaps 3 months of that towards the end. I was on the call with dozens of other people whilst me and one of the other guys hooked the final pieces together and tested it end to end for the first time... and it worked. There was actual cheering.
I had a similar experience where someone that had prior experience with django while we were using sqlalchemy started to design a django-like ORM on top of sqlalchemy.
Of course it took him some time to get working, was a hell to understand and extend and most importantly lacked support for basic features.
Fortunately it was limited to a small isolated service but I can imagine the damage on the long term if you continue that route on what becomes a huge monolith after a few years.
I was complaining about SQLAlchemy's insane quirks to a group from my alma mater and one of the grad students said, "Well, the solution to your problem is clear: Write your own ORM." and I had to explain that this startup does not want to get into the ORM-writing business.
Also, some people want to work on what's already familiar to them. If building a framework from scratch is what appeals to them, they'll do that even if a framework already exists. Busywork to look productive.
I resonate strongly with this story. I’ve seen three people teams get in one month where SAP could not in year, and also let and witnessed incredible number of total fakers in SaaS enterprises.
Big corpo may be too big to fail, not so sure about their whole cohort of partners and fakers.
We are certainly closer now to being able to prototype and go to market faster with a product. In one weekend is a little much but I think its hard to deny that building will continue to expedite. What most developers don't think about is that the marketing, sales, customer service are all non-trivial parts of the business/product and all require legwork that is more than just sitting at an IDE. The nail in the coffin is that the data is a large part of company moats, and new products need time in the market to get that. Migration is also a long process and risky...so to get customers, a newcomer needs to provide way more value than what the incumbent gives.
I imagine you're going to have people trying to automate the whole GTM lifecycle, but eventually the developer that thinks they can bootstrap a one man enterprise without actually doing any kind of social interaction will run into a wall.
I sure hope I never have to hunt down any GTM options myself soon and I can tell the AI to do what it should be doing. However AI adoption may be getting slowed down by profit motives because what Google should have already been doing is letting me git clone the entirety of GTM with all its configs to a local folder so I can treat it like code because it is code. The difficulty with AI adoption will be to make all products be like this so they can interact on a code level instead of me having to press buttons in different UIs to make thing shappen. E.g cloudflare should be letting me git clone its entire config, everything I did in the dashboard, to my folder too.
It's shocking to me how prevalent this "who needs Salesforce when everyone can just vibe code their own CRM from scratch in a day" narrative has become in the business press. Like, what???
I don't think vibe-coding will replace anything. However, what if AI tools can make skilled developers more productive, particularly at simple tasks in unfamiliar environments? You could see that reducing the engineering costs of simple utility applications. There are tons of pitfalls that many here have pointed to but also maybe opportunities to do things that wouldn't have been cost effective.
Also: In my life the easier it has gotten to create and run software, the more software people have wanted and the more they have been willing to spend on it.
People really seem to believe that code is the only thing you need to make a SaaS company. It's like thinking a line cook is all you need to open a restaurant. There are so, so many other components to running a business.
I don't think it will replace SaaS but I do think it can replace the need for a lot of the consultant work that goes around configuring and integrating the SaaS. It will be much easier to have a spec that defines how things need to be configured and the machines can implement it (using the SaaS as tools). Frankly this is the most annoying part. It's not that the B2B stuff can't do whatever, it's that it never gets implemented in ways that aren't a pain in the ass because it's all handled by people who aren't actually using them.
I really don't think it's not going to become "these prompts are specs" and then you have processes of reviewing implementations. It's one thing when you have randos building stuff and they leave etc. Having stored prompts and managed code that uses tools is a different beast.
No serious programmer "vibe" codes. I admit creating SaaS may not be feasible with current infrastructure but you can't ignore the insane jump in productivity that these tools can offer with the right scaffolding.
What I struggle with is developers wanting to leave platforms like Datadog for open source equivalents that need to be self-hosted.
I hear all of the cost savings benefit, but I never see the team factoring in their own time (and others time) needed to set up and maintain these systems reliably long term.
Something IC’s at company often struggle to understand is the reason why companies often prefer to buy managed solutions even when “free” alternatives exist (read: the free alternatives are also expensive, just a different type of cost)
My log bill for Google cloud log would be like 30k. For splunk I like 80k. I self host for 1.5k per month. Spend maybe an hour a month? Easiest money I ever made.
Do they actually not understand that? They might just be fine with a system that makes them more useful.
How do you calculate the time spent on an internal tool like this, actually? (I’ve never been in management). Realistically your team inevitably will have some downtime, maybe some internal tool maintenance can be fit in there? I mean it obviously isn’t fully “free” but is also shouldn’t be “billed” at their full salary, right?
>the free alternatives are also expensive, just a different type of cost
Not if you hire reasonably competent people. These days for vast majority of FOSS services all you need is an ability to spin up a VPS and run a number of simple Docker/Podman Compose commands, it can't be that hard.
I'm sorry but the amount of companies that need something like DataDog is quite small compared to their 30,000+ customer count. Maybe 5,000 companies on Earth truly need something like DataDog, 80% of their customers would be perfectly fine with a self hosted instance of grafana.
Using an open source self hosted solution should be the industry standard, encouraged position, by default. Our industry does not gain overall from using DataDog but only from truly open source solutions that utilized AGPL licenses that allows everyone to move forward together + share lessons together + contribute together toward a common goal of better observability.
Why are we acting like it's hard to set up? This isn't the 1990s, it's 2026. Tooling has gotten quite good over the last decade.
Also corporations stupidly spend money all the time, they over spend too. I recently left a company that was paying SalesForce $10mil a year in licenses when only 8 people in the entire 3,000 person company was using it. I doubt that was the only single instance across our industry too. There is a massive amount of waste and graft in enterprise sales.
I honestly doubt it if you replaced grafana for 10,000 DataDog customers they would notice the difference.
& the counter-argument is those SAAS apps being killed by A.I are growing revenue 20%+ YOY
people who write this BS - one never don't understand SAAS fundamentals, they only see what's on the screen and forget the complexity that lives on the backend - forget the costs of running such a SAAS
before it was low-code will kill SAAS, then Visual UI builders, now its A.I
just like it was before that crypto will kill Trad-Fi
people who say these things - have tied their identity into it so they whole-heartedly believe the bullshit they say even though reality doesn't match
to anyone curious read the 10k (Annual Report) of any public SAAS - Salesforce | Workday etc - people should admire these companies for the machines / ecosystem they built - and also learn the good & mistakes to avoid i.e the bad
those annual reports tell you how the revenue generation machine works, how much revenue is expected 2+ / 3+ years from now - their weaknesses | headwinds and also tailwinds - how those companies grow and continue to grow etc
SaaS companies are effectively both "software maintainers" and "price gouging middlemen" at the same time. The difference between the bid and the ask for SaaS is part of a simple math problem for whether the company should try to create their own version of the software they need. It may be the right decision, it may be the wrong decision, but it will be the right decision for a non-trivial number of firms. And that means SaaS businesses will both lose customers and have downward pressure on their margins. That means valuations of B2B SaaS firms go down.
My career occupies a weird middle ground where, for 20 years or so, I've catered to smaller businesses that need bespoke solutions (because the SaaS available doesn't conform well to their business logic), but don't have the scale or desire to build and maintain software in-house. Sometimes these are slapped together in a weekend, if that's all that's needed. But in most cases they still become ongoing improvement and maintenance projects for me.
This niche position has had some interesting ramifications for them and for me. They clearly incur a lot of technical debt once their business relies on bespoke software. On the other hand, they own the software and can get an immediate response or new feature or upgrade from me, limited only by my time. And in the end, this ends up saving them time and money. It gives me a permanent and unending flow of work. But if I die, they're pretty screwed.
One reason I don't vibe code things even now, even simple components that could easily be vibe coded, is that I remember and know where everything is, every function or line of code that might be causing issues, because I wrote it myself. I know right away where to look for a query that might be throwing errors after a database upgrade, for instance.
As a manager I assume you would probably not want to go down the road of hiring someone like that, but for companies of a certain size it's an acceptable compromise. However, I wouldn't want to hire someone like that myself unless they were extremely reliable and didn't rely on AI to write any of their code.
People sometimes fail to appreciate the value of KNOWING the system inside and out when it comes to diagnosis and troubleshooting.
Observability is great, dont get me wrong, but past 3 to 6 months of work on the same thing...I can almost beet the observability tools in timetoresolve.
This sounds great if you get on well with your clients. You must be an effective networker and at sales. How do you bill, and how do you price your services?
It seems that current advantages would compound with AI. I.e., if I am making a SaaS for Popsicle stick makers today, why I am disadvantaged with AI vs a new competitor in the space? I guess the hypothesis is the Popsicle stick maker will vibe code all of the software that they need instead. For that, we need significantly better AI than we have today - perhaps something like a 1000X improvement. Basically, this is a world in which non-technical grandparents can vibe code anything that they want. This means, it understands what you want without you being able to articulate it well in the first place.
I don't do tea leaves so I wouldn't commit to that, particularly because I think SAAS was oversold in general even before LLMs came out. But I think the idea that the industry as a whole will shrivel away just isn't feasible, even if there is a correction.
I dont even blame management. I believe most of them are well-aware that much of what is going on right now is pure hype.
However, they dont have a choice. The sentiment of shareholders is that they want their cash (yes it is their cash that managers re-invest on their behalf) to be invested in AI-related projects.
So...... you get what you get, and investors will get what they deserve. But they will still blame the management in the end ;)
All of the hype surrounding AI will subside when a SaaS company eventually deploys a moltbot version of their software and the company is driven out of existence due to the chaos that ensues.
I've seen some bad some SaaS that management insists is an integration they need and no one else provides. I can easily see some vibe coded projects replace them.
They will magically realize this when their huge bonuses will be tied to something longer lasting than last quarter/year performance on some very narrow metric (which has nothing to do with sane stuff like adding long term value to some part of the company).
They are not stupid, far from it, most are (very) high functioning sociopaths. And out and up there its everybody for themselves first.
A huuuuuge part of why companies, especially public companies, especially those in regulated industries like healthcare and finance are willing to pay eye-watering sums of money for a SaaS app that you could get an MVP up in a few weeks time from a competent team with no AI is that they need a phone number to call when something shits the bed at 2 AM on a Wednesday. They need support SLAs without the payroll and headache associated with it. They need someone to sue if things truly go tits up.
Moving SaaS apps in house is a great way for a VP to get a fat bonus or a director to get a promotion but I have to imagine it keeps the CIO/CTO up at night unless they're fully asleep at the switch.
Yep! We sometimes have a choice between the gold-standard and commonly updated open source solution to X and a two-bit hacked together proprietary solution that has 24/7 support at high cost...and we choose the one with support, because that's what our audits basically require. Because then we can say "yes, it's still within the support contract, we have an escalation point".
In my career I've deployed close to 100 different saas products at enterprise level and can tell you that most of the current crop are slapped-together half-finished dross with a huge sales and marketing team.
In the time it takes to deploy semi-bespoke saas, or while waiting for the current licence term to expire it would be very easy to develop a more suitable and much cheaper product in-house, this was true before AI tools and doubly so now.
The problem is right now management is not only insisting on their team vibe-coding bespoke replacements, they’re avoiding paying for other SaaS because they can vibe-code their own replacements, often themselves, and they’ve lost sight of that they probably don’t want to be responsible for it.
Besides rampant failures in communication and skills allocation, wild U turns of requirements were (sometimes, not even real business requirements) were holding back corporate environments doing a decent job.
With AI, I can only see the rate of such changes sky rocketing due to expectations wildly misaligned with reality. Hence we are unlikely to see any meaningful improvements.
This is an incredibly broad statement that just isn’t true in a million cases. Last co we migrated all of our observability from Datadog to Grafana/AMP because it’s much much cheaper. The vendor can charge some premium over the cost to build/maintain but not infinity. SaaS is going to have to get dramatically cheaper to compete with the lower cost of building your own.
You're not considering opportunity costs and buyers vs. users.
If your senior developers can slap together something better than an expensive SAAS offering you want them directing that energy at your core products/services rather than supporting tools.
And the people deciding to buy the expensive SAAS tools are often not the people using them, and typically don't care too much about how crappy the tool may or may not be for doing the job it's advertising as doing.
Clubbing all saas products together just means you can’t really have a productive discussion. Saas products are on a spectrum of quality, from amazing (stripe, datadog) to terrible (fivetran, github). Its upto you as a user to make a call as to which will serve you best, what you should focus your limited resources on etc.
> what if this time it's senior developers and they actually can slap something together better then the expensive SAAS offerings
A typical SaaS customer will use many pieces of software (we mostly call them SaaS now) across its various functions: HR, accounting, CRM, etc. Each one of those will have access to the same pool of senior devs and AI tools, but they will pour more resources into each area and theoretically deliver better software.
The bigger issue here is the economics of the C-suite have not changed here. Assume a 100 CPG company uses 10-20 SaaS apps. Salesforce might be $100k/year or whatever. 1Password is $10k. Asana $10k. etc. They add up, but on the other hand it is not productive to task a $150k employee with rebuilding a $10k tool. And even with AI, it would take a lot of effort to make something that will satisfy a team accustomed to any modern SaaS tool like Salesforce or Atlassian. (Engineers will not even move off Github, and it's literally built on free software.)
That's before I get to sensitive areas. Do you want to use a vibe-coded accounting system? Inventory system? Payroll? You can lose money, employees, and customer perception very rapidly due to some bugs. Who wants to be responsible for all their employee passwords are compromised because they wanted to save $800/mo?
Then, the gains from cutting SaaS are capped. You can only cut your SaaS spend to zero. On the other hand, if you have those engineers you can point them at niche problems in your business niche (which you know better than anyone) and create conditions for your business to grow faster. The returns from this are uncapped.
TL;DR; it's generally not a great idea to build in-house unless your requirements are essentially bespoke.
To the first question, if your senior devs can do that there's almost certainly something more directly valuable to your business they could be doing than solving a problem your vendor has already solved
The second question is a valid one, and I think it will somewhat raise the bar of what successful SAAS vendors will have to offer in coming years
Nice what ifs, but not valid so far. I get the motivation to think/hope so, but thats not the proper business world right now where big money are. Maybe next year it could start becoming true but then market will be a bit different too
There are of course exceptions to every rule, and I'm sure some companies have been successful in building their own in-house tooling.
At the end of the day these decisions are all series of trade-offs, and the trick is understanding your requirements and capabilities well enough to make the right trade-offs.
I think there’s also the classic “I can build zoom in a day” - they get video working between two machines. But it’s the last 80% of the app that takes 99% of the time. Nerds don’t see the whole product, just the happy path of a wee technical challenge.
This may be true pre-LLMs, but I think you need to account for the baseline build-vs-buy tradeoff shifting.
Companies in most cases don’t want to build SaaS because it is expensive to hire engineers to do it, not because they are allergic to owning teams.
If in-housing becomes substantially cheaper than the alternatives then companies will adapt.
But even if the new equilibrium is to hire a contract dev shop to build something custom to keep avoiding responsibility, this would have the same impact on SaaS.
So I’m pretty skeptical of this first-principles prediction expressing right level of uncertainty.
I totally agree about the management reluctance to just own everything in house.
But I think it’s plausible that SaaS companies will be easier to start with AI coding, and with lower costs (thanks to AI) they will be able to get into the black with a smaller addressable market. So each one can have a different mix of fewer features, for different segments of customers, at lower prices.
The result would be a loss of pricing power by the incumbent do-everything big guys: no more baked-in 10% annual increases. Which is still a pretty big change in their economics. And therefore valuations.
The companies that already have a strong in-house team will greatly benefit from AI. Many of those who don't are in that situation because managers have PTSD from so many failed projects.
Half of all projects fail. That's a lot of emotional trauma.
This was all possible pre-AI. The reasons that some Saas companies win have nothing to do with how quickly or cheaply code can be written for the Saas.
I can't count the times I've told clients and prospects to _not_ hire us to build something they wanted. Because they could just use off the shelf solutions that were cheaper financially, at least in the short to mid term, and much, much cheaper in terms of opportunity costs. I struggle to put even billed hours into something that doesn't make sense to me.
Of course some overdo it. I've seen companies with more random SaaS tools than staff, connected with shaky Zapier workflows, manual processes, or not at all. No backups, no sense of risks, just YOLOing. That's OK in some cases, in others really not.
I suppose it does need some engineering thinking to find the right things and employ them in a good way. Unfortunately even developers commonly lack that.
I'm a manager too, but I'm also the new guy pushing the solution to a human problem: work management. SMAR, ASAN, MNDY, etc. Not only do people not want to be responsible for it (and in some cases simply be "not responsible"), not only is the internal solution "too time consuming", the only answer thus lands on hiring external consultants to implement and maintain massively-overkill-$olution$ in $aa$ like CRM, NOW, etc. which as you know, do not solve the same problems as the aforementioned SaaS.
"Now that I'm in management, I 100% get it."
100% and win or lose I am still going to fight it...
But this time management has to justify its AI spend.
We've been through cycles of outsourcing and in-housing skill before. This seems similar but for tools and services. Maybe we'll see internal teams taking the reigns back to replace bad-fit SaaS products.
There's still a lot of risk associated with in-housing though (perhaps more than before). That means the real opportunity is for new, leaner B2B SaaS businesses to step in, especially if there's a displacement effect from seeing internally built prototypes of expensive subscription software.
There's that and then there are companies spending 100k on a software suite just to use that 2 features. So now one of their junior dev solves it and becomes a hero. The truth it always somewhere in the middle.
I think the pressure on SaaS margins won't be from customers vibing their way to Figma or DataDog but because gen AI will bring a lot more credible competition in many segments. DD and Figma are probably awful examples because those companies are constantly pushing the envelope, but there are a lot of rent seeking SaaS providers that are going to be in for a rude awaking.
The difference between a vibe-coded prototype product, even a good one, and an enterprise SaaS platform is the difference between a Lightning bug and a Lightning Bolt.
To be fair, re-creating the SaaS solution that simply replicates the features they see can often be done fairly simply. However, there are generally a whole lot of things under the surface. Then there is the whole hosting and maintaining the system, which is its own problem.
I've seen this happen with both juniors and seniors. They do come back with a working solution /for the happy path/. Because the happy path is easy. It turns out that most of the complexity sits in the unhappy paths.
I agree. Just because you can buy some piece of software doesn't mean you should -- there is a lot of software that exists just to sell more consulting hours and will never fit the business. It's actually not hard at all to code and maintain much simpler alternatives.
Actually having to support multiple businesses with commercial software is hard. I've written a ton of custom software that far surpasses the capabilities of commercial offerings but if were to turn that into it's own commercial offering it would be large undertaking.
If the management is the one actually paying for the software from their own pocket (founder), the tables turn. There are millions of SME owners who are forced to pay for B2B software just out of necessity and not having resources to build it in-house.
I work at a smallish public tech company and while this may be true at some companies, it's not true at all of them. We have almost no SaaS vendors. If we do have to buy software, we're almost only interested in On-Prem.
It almost always devolves into some all encompassing ERP that is meant to solve the needs of all parts of the business and save millions in licensing costs, and we all know how well that plan goes.
Software without support is useless. In the business world, what's being bought isn't code—it's a solution to a business problem, with a throat to choke if things go wrong.
management simply doesn't want to be responsible for it
That sounds dysfunctional. The purpose of management is to manage risk, not to avoid it. A proper manager would be able to quantify both the risks and the costs, present those figures in an easy overview, and then be able to defend their decision (or advise higher management) using that.
My firm has partially transitioned through this curve. We went went from "fully externally supplied" systems, to an architecture that combines "externally supplied" (core functionality) with "low code" about 6 years ago. I would argue (as a financial manager) that that lead to a more flexible and more affordable architecture. A funny mixed bag problem arose though: the curve of business demands grew harder than that of IT-delivery. So IT delivered more value, but business keeps demanding a faster pace. If I project this line to the future AI will most certainly harm our external suppliers. We keep getting better at DIY development and "low code" will transition towards "no code". Not really "no code" of course, but DIY IT developed tooling.
The age of the business developer has re-arrived.
For the first time in my career, I can point to multi million euro external suppliers, tell my environment "that's basicly an API + authentication from X to Z, let's develop that ourselves" and get a response of "When" instead of "No". B2B SaaS is toast in my perspective, as are boutique firms delivering solutions + consulting. I can create a million euro team easily (that's like five developer years), if they deliver a successful insourcing. And now I feel like writing MBA-slop, but's it's all about growing your IT maturity. All insourced code is future maintenance expenditure. You need to balance that to the benefits.
> All insourced code is future maintenance expenditure. You need to balance that to the benefits.
I love this perspective. I feel like the pendulum has swung too far back to "it's easy to build, it'll be easy to support". But to be fair, it was probably too far the other way a few years ago: "it was easy to buy, it'll be easy to have them support it".
Other than trial and error, how do you think about pricing out maintenance costs for insourced code vs purchased functionality?
> management simply doesn't want to be responsible for it
The problem with this kind of thinking is that it strips away all nuance. At some point you have to be responsible for something ... otherwise you don't have a business. You are simply a wrapper around your SaaS providers and tightly coupled to their success. The key is knowing when to offload and when to keep it in house. Quite frankly, your average weekend MBA VP simply doesn't have the expertise to make these kinds of decisions. This is why so many VPs exit before things get bad.
Think about it differently - let's say a free OSS product can be installed and you can use ALL features except for LDAP (because that's the paywalled portion that requires you to buy it for $25k / month.)
Well, with claude, you can download the code, tell it to implement LDAP authentication, and smile all the way to the bank. And for said fortune 500 company, employing an employee to spend 100% of their time maintaining the app at 10k per month is a 15k savings! And because it _doesn't really take 100% of their time_ it's really only like $500 per month? And to be completely honest, how man times did you get Jira to fast-track your issue?
I get it however, the manager angle. It's still a distraction. But the article being referenced still shows revenue going down.
There's definitely a lot of cope in here, mostly because SaaS is keeping them employed... be ready, the crush is "almosthere".
If I understand correctly many organizations will not develop original stuff internally, because nobody internally wants to be the one is shouted at if something goes wrong.
I like to bring up JIRA example. You could replace it in-house yeah it is just tickets with statuses. /s
But then keep in mind one who built the replacement will become the owner of an application that business doesn’t want to pay for and that person will be cost center for the company.
That person better get marketing and negotiating skills that Atlassian has on board because that person will be responsible for the app and will not be getting salary increases for working on something that is not core business of the company.
You guys keep using services like Jira, Salesforce, Stripe, Datadog, etc. While those are definitely the biggest names, I don't think people are referring to those SaaS platforms as the ones they will replace or try to build an inhouse version of. It will be things like ETL pipeline services, data scraping services, maybe some internal analytics SaaS. The niche things that cost a lot because they’re in a sweet spot where only a few people need them, but no one used to have the resources to build them in-house. So, when the salesperson called and offered a perfect solution to their problem, they bought the service. Those are the ones that will be more targeted for in-house solutions.
I'd actually say the opposite is the case. B2B (even SaaS) is probably the most robust when it comes to AI resistance. The described "in house vibe coded SaaS replacement" does not mirror my experience in B2B at all. The B2B software mindset I've encountered the most is "We'll pay you so we don't have to wrestle with this and can focus on what we do. We'll pay you even more if we worry even less." which is basically the opposite of...let's have someone inhouse vibe code and push to production. B2B is usually fairly conservative.
There was no chance that everyone would be running their own email server, but if it wasn't for the lack of IPv6 adaptation a plug and go home email server solution would probably see a decent amount of use. I'd bet we'd already be seeing it as a feature in most mid-ranged home routers by now.
What a weird take. I was running my own email server 25 years ago on a 512 kbit ADSL line. No problem at all, would even be enough bandwidth today for most messages.
(Back then email still worked from residential IP addresses, and wasn't blocked by default)
I agree with you. In B2B SaaS you don't sell the software, you sell your expertise in a specific domain and the responsability you take for owning that expertise. The fact that the development costs are nearly zero will make them more valuable and more protifable
My experience is that SMBs are generally not run by people who feel confident doing any kind of self managed IT.
No amount of LLM usage is going to change them into full stack vibe coders who moonlight as sysadmins. I just don't see it happening.
Not until, that is, a new generation, that has grown accustomed to the tech, takes over.
Until then the current SMBs will for the most part fulfill their IT needs from SaaS businesses (of which I think there will be more due to LLMs lowering the barrier for those of us who feel confident in our coding and sysadmin skills already).
I'm considering SaaS replacements with in house code in situations where my general thoughts are "how can this possibly be the pricing for this?" which is not uncommon.
Well before vibe coding, tons of open source software existed (and exists) to replace SaaS. With lots of features and knobs and real communities. But I still often pay for SaaS because managing it is a headache. Some human has to do it. I can pay the human or I can pay the company. I really don’t see how vibe coded toys can replace real battle tested SaaS products. A better explanation is the bubble in PE ratio is deflating and it’s happening all over, regressing to the mean. AI is a convenient explanation for everything
Maybe you are right and the companies do want to pay and not worry about these problems. But now they have a lot more SaaS options to chose from. The incumbent companies like Salesforce and Atlassian have less of a moat. Maybe they'll keep the power users but if a customer is only using 80% of the feature set there is new competition.
Competition might come in the form of a startup but it can also come from existing SaaS companies expanding into adjacent domains. Canva now does docs. Notion does email. etc
Also, it is my experience that exec and boards favour safe and well known B2B partners over in house. It's a more publicly defensible approach that gives them an out if things go wrong and shareholders get upset.
For big corporations at least prices of SaaS are rarely an issue. Issues are: we don’t have the time to introduce a new tool, what about our processes, we don’t have the right people.
> we want recent examples just look at tailwindui since it's technically a SaaS.
This is a terrible example. Show me someone ripping out their SAP ERP or SalesForce CRM system where they're paying $100k+ for a vibe coded alternative and I'll believe this overall sentiment.
TailwindUI isn't really what I'd consider SaaS -- it was a buy once and download software product.
That means to keep making money they need keep selling new people. According to them, their only marketing channel was the Tailwind docs, AI made it so not nearly as many people needed to visit the tailwind docs.
If they had gone with the subscription SaaS model, they'd probably be a little better off, as they would have still had revenue coming in from their existing users.
> I mean if we want recent examples just look at tailwindui since it's technically a SaaS.
How is it in any way B2B? At most B2C + freelancers / individuals / really small SME.
It didn't have any clues a med/large B2B would look for e.g. SSO, SOC2 and other security measures. It doesn't target reusability that I as a B would want. The provided blocks never work together. There aren't reusable components.
Tailwind UI or now Tailwind Plus is more like vibe coding pre-AI.
Sorry but tailwindui is not a SAAS. There is no service or hosting. You buy a coded template once and then receive updates. It is totally not the same as a critical B2B SAAS that is running 24-7 on the vendor's servers providing real support and service.
TailwindUI unfortunately sits in a position of being an easy to disrupt business with current AI.
Now attempt the same with Zoom, I suspect vibe coding will fall down on a project that complex to fit the mental model of a single engineer maintained a widely used tool
Perhaps the case for premium CSS SaaS businesses, I guess (which seems particularly primed for disruption even pre-AI), but there are many more robust B2B categories out there that aren't literal code + docs as a service.
how dont people understand? if you have a VC funded b2b saas, you need to charge huge margins for the investors to get a return. now, small teams can vibe code a replacement and charge 90% less money. AI is going to kill saas margins.
i literally cannot understand why people keep repeating that non tech companies will build their own software, thats not the bear case for saas
Yeah.... The code isn't the hard part. That's not where the value is.
This hard part when you're doing in house stuff is getting a good spec, ongoing support, and long term maintenance.
I've gone trough development of a module with a stakeholder, got a whole spec, confirmed it, coded it, launched it, and was then told it didn't work at all like what they needed. It was literally what they told me... I've said 'yes we can make that report, what specific fields do you need' and gotten blank stares.
Even if you're lucky and the original stakeholder and the code are on the same page, as soon as you get a coworkers 'wouldnt it be nice if...' you're going to have a bad day if it's hand coded, vibecoded, or outsourced...
This has always been the problem, it's why no-code never _really_ worked, even if the tech was perfectly functional.
The accounting saas dores presumably uses doesn't "automate spreadsheets" as its core value prop.
related: i'm thinking these vibe coded solutions are revealing to everyone how important and under appreciated good UX is when it comes to implicit education of any given thing. Like given this complex process, the UX is holding your hand while educating you through a workflow. this stuff is part of software engineering yet it isn't "code".
I, on the other hand, can't wait to fire every single B2B subscription we've got.
B2B SaaS is a VULN. They get bought out, raise prices, fail. And then you have extremely large amounts of unplanned spend and engineering to get around them.
I remember when we replaced the feature flags and metrics dashboards with SignalFX and LaunchDarkly. Both of those went sour. SignalFx got bought out and quadrupled their insane prices. LaunchDarkly promised the moon, but their product worked worse than our in-house system and we spent nearly a year with a couple of dedicated headcount engineering workarounds.
Atlassian, you name it - it's all got to go.
I just wish I could include AWS in this list. Compute and infra needs to be as generic as water.
If you're working at SaaS, find an exit. AI is coming for you. Now's a great time to work on the AI replacement of your product.
> And then you have extremely large amounts of unplanned spend and engineering to get around them.
I have no idea how you are spending "large amounts" of unplanned spend on Saas products. Every company I worked for had Saas subscription costs being under 1% of capex. Unless you add AWS, which is actually "large amounts" but good luck vibe coding that.
1. This isn't rooted in data but anecdotes "One Series E CEO told me that they’re re-evaluating the quarterly renewal of their engineering productivity software because they along with an engineer reimplemented something using Github and Notion APIs. They were paying $30,000 to a popular tool3 and they were not going to renew anymore."
2. These anecdotes are about tech startups spend, not your <insert average manufacturing business>. Nor or they grounded in data that says "we interviewed 150 SMB companies and 40% of them have cancelled their SaaS subscriptions and replaced it with vibe coded tools"
- If our customers vibe coded better integration points for us, it probably improves our overall value to our customers.
- The software industry, especially startups, is such an insignificant portion of the market, its not really worth worrying about. But, I can tell you from experience, that even large software companies don't want their own developers spending much time on accounting, ERP, or HRIS systems and they "outsource" this to SaaS companies.
Yeah, this is just long-form linkedin slop. He's thought-leadering to get you to get his (no-doubt slop-written) guides and do leadgen for his forthcoming saas.
Yeah, also if a SaaS costs, 10k a year, I promise its not not more cost effecient to pull your 10k a month engineer off their usual work to build and then maintain some vibe coded slope everytime an edge case occurs.
Also many customers of SaaS have little to zero engineering staff, they are in construction, resturaunts, law offices ect. These takes are so assanine.
Is there a place on the internet where folks like yourself, who seemingly have a way to think economically congregate? I personally dont know of one, for which if I did, I wouldnt visit here anymore.
So many takes on here are so lazy and simpleton that when you go a few levels deeper all the flaws get exposed.
Even in companies that have SWE, do you really want to divert in-house SWE time to something as exciting as ... accounting rules and making sure your inventory is auditable? Or any number of the weird compliance things associated with most B2B software for a medium-size business?
I have few sticks in the sand in my thinking framework:
* writing code has always been the easiest part of building software, deciding what to do and what not to do is something else that takes forever sometimes
* there are several open source projects that can replace commercial SaaS and still people prefer to purchase commercial SaaS. These are available immediately, deployed immediately etc etc.
* along the same line, some of those open source projects offer self-hosting and cloud version: I would always personally go for the cloud version because in a small team I don't want to operate something that other people built and I don't know how to operate. That's not my job not my team job
* people are underestimating how draining is operating and maintaining software, which is something that goes beyond the adrenaline rush you get after "building" something with Lovable or similar tools. Also, I find it extremely easy to get 80% done quickly but excruciatingly slow to get things done right.
* I still see huge value in using tools like Lovable to build a working prototype and validate assumptions so that you get quickly build the right thing right solving the right problem in the right away avoiding waste
* camcorders have been around for ages but you don't have millions of directors around just because you make a tool more accessible
* same can be said for other things like restaurants, where sometimes it's more convenient (although expensive) to buy vs build.
>people are underestimating how draining is operating and maintaining software
Yep. Many SaaS have an edge because they factorize the struggle of many customers, if a SaaS has 1000 customers, each customer vibing their way into a home-built solution will require dedicated efforts at maintaining it. Even with AI, those efforts aren't negligible.
Many companies don't even operate any IT infrastructure, cloud or otherwise, themselves, beyond office connectivity, AI replacing SaaS will require someone in charge of that at the very least.
> camcorders have been around for ages but you don't have millions of directors around just because you make a tool more accessible
tiktok alone has 1.5 million directors! it's just that we call them creators now.
the meaning of the word director has changed, that's all. but professional roles shift in meaning all the time. a computer used to be a literal human doing arithmetic. an engineer was someone who designed war machinery. being a doctor used to mean teaching at an university.
human beings are natural tool makers. we always have been. the frontier material to manipulate where the most advanced engineering happens constantly changes: stone -> bronze -> iron -> ink (descartes) -> steel -> silicon -> javascript (YOU ARE HERE).
notice how each step is an improvement/abstraction on top of the steps that came before it. some say english is next in that chain. i honestly have no idea. all i know is there will always be The Next Thing and it'll be much nicer to work with.
> camcorders have been around for ages but you don't have millions of directors around just because you make a tool more accessible
That’s exactly what happened. There are more filmmakers than ever now due to the accessibility of cheap cameras, then digital tools, including affordable HD cameras. Especially once the DSLR revolution took off circa 2011, which enabled budget-constrained aspiring filmmakers to use prime lens sets rather than fixed/built in zooms on cheap camcorders. With proper lighting they could actually make something look pretty damn cinematic. The entire industry has radically shifted in the last 15 years in particular due to these changes, but it started to shift around 2000 if I have to assign a particular year to it.
When you had to shoot on film stock, which was expensive and had a whole processing pipeline that one couldn’t reasonably do at home, there was a much thicker barrier to entry. You basically had to go to film school or get into the industry before you could start making your own stuff. Hell the Duplass brothers started out on crappy camcorders. Now? A smartphone, some cheap LED’s, a basic computer, you can really make something.
My point was along this line: writing code is different than building a product as much as recording a video is different than telling a story with a movie (long and short that you want).
Do everyone has the capability to build a comprehensive set of features that we call a product to solve problems that people or business have in their life?
That’s why I’m always skeptical about measuring AI impact based solely on quantitative metrics.
Let me provide some possible evidence against this: so many teams are desperate to rewrite their codebase but struggle to actually do so. And when they finally make the leap, it takes them 5x as long as they had hoped. Then sometimes the new code isn't even any easier than the old code.
I personally find writing code to be a huge time suck and I'm very happy that AI helps me save that time.
But again, how much of that time is spent writing code?
I've done rewrites, replatforms, a bunch of times. The actual programming is not the tricky part, but instead (1) picking apart the legacy system, understanding what to build, (2) orchestrating the work to shift transactions from service A to service B without breaking anything.
Teams and especially developers love to think they can skip that phase and just crack on with the programming, invariably what happens is the same as described above: intoxicating velocity followed by a hard stop when you realise you haven't solved problems (1) or (2) above
Most of a software project's lifetime will be spent as a maintenance challenge. i.e. How do we add the 237th feature without adding to the performance problems that already exist. Hence, the desire to rewrite the codebase to incorporate the abstractions of all 236 features.
I don't see AI helping with this. From my experience, it seems like the opposite. It can help you write the code after you've deconstructed the problem yourself and know how to keep it in check.
I reckon a lot of re-writes tend to take "them 5x as long as they had hoped" and "isn't even any easier than the old code" exactly because writing the code wasn't the problem in the first place.
It's business logic, edge cases and other small necessary details that accumulated over time which make the code 'messy'. Once you've integrated all those in the new system, it likely looks equally messy. And discovering and implementing all those extra requirements is probably what took you the longest.
Not to say this applies to all re-writes or that AI tools can't help the process
> “…deciding what to do and what not to do is something else that takes forever…”
If that’s the case, then I should think business owners and office workers should be able to sort that out, lestwise on the “how to automate the boring stuff” front. That repetitive, boring, time consuming, error prone work. Incremental, least work for greatest impact.
The danger is they pull a 1999 Mars Climate Orbiter —level mistake. Or their solution suite grows to big to manage with mounting tech debt.
Also, if you’re software also required some custom domain hardware, then there’s your bottleneck for protectionist business practices.
I see that Software as a Service banked too much on the first S, Software. But really customers want the second S, the Service.
When you sell a service, it's opaque, customer don't really care how it is produced. They want things done for them.
AI isn't killing SaaS, it's shifting it to second S.
Customers don't care how the service is implemented, they care about it's quality, availability, price, etc.
Service providers do care about the first S, software makes servicing so much more scalable. You define the service once and then enable it to happen again and again.
They didnt, dont make the mistake of thinking Saas companies are just software companies. They are Sales companies who happen to sell software. Companies like Dropbox & Atlassian have long been surpassed in Tech but they live only because they continue selling even when demand was hard to get. Their moat is sales & networking and software has to be just good enough. And other part is service, these companies still have one of best costumer service since the start of early 2010s. You can still get refund on Uber quite easily, but if you try doing that at a regular old school company you would require a prayer and couple of business weeks.
Yes, many don't like Sharepoint, but still they use it. It's the tool they can use.
Customers don't care if Sharepoint uses LLM, they just want to share ideas, files, reports, pages, etc. If LLM makes it easier, great! If some other product makes it easier, great!
It's not about the product it's about the results.
You're proving the point? Sharepoint, teams: availability + price. Every company has microflows, sharepoint and teams are automatically available and part of the price or lower priced than the competition.
Nah it's not that at all. Most of the services are totally fungible and everyone has a short attention span. You need to be in a market which is extremely difficult to disrupt and have a product which people are totally dependent on. And those tend to have a rather large cost to enter unless you were in early.
I just don't want to pay $50/user/month for an initially open source product that was relicensed and then crippled that the initial group giving something away decided they wanted to make a business of it.
I think one of the interesting things here is that AI doesn't need to be able build B2B SaaS to kill it. So much of the overhead of B2B SaaS companies is thinking about multitenancy, intergrating with many auth providers and mapping those concepts to the program's user system, juggling 100 features when any given customer only needs 10 of them, creating PLG upsell flows to optimize conversions, instrumenting A/B tests etc...
A given company or enterprise does not have to vibe code all this, they just need to make the 10 features with the SLA they actually care about, directly driven off the systems they care about integrating with. And that new, tight, piece of software ends up being much more fit for purpose with full control of new features given to company deploying it. While this was always the case (buy vs build), AI changes the CapEx/OpEX for the build case.
And in many cases, it's 12 features, with 2 of the features not even existing in the big SaaS.
I'm pretty sure every developer who has dealt with janky workflows in products like Jira has planned out their own version that fits like a glove, "if only I had more time".
If companies wanted to build thier own simple-JIRA they could have built themselves before. I dont think making a kanban board was hard even before AI.
JIRA especially, and I'm always shaking my fist at Atlassian that simple APIs or workflows or reports aren't already included in the tool. I have to pay some other company $10/user/month to get this dumb report your tool should already be able to do?? Insane.
Jira has had free competitors that do at least 75% of what it does since it's inception. You could find a dozen on github that actually look good right now.
Until a given company decides they need access control for their contractors that's different from their employees, etc. etc. etc. - seen it all before with internal often data scientist written applications that they then try to scale out and run into the security nightmare and lack of support internally for developing and taking forward. Usually these things fizzle out when someone leaves and it stops working.
Pretty much. My employer was looking to cut costs and they were spending ~500k a year on a product that does little more than map entra roles/groups to datasets and integrated with a federated query engine through a plugin. Took a couple days to build a replacement. The product had only a few features we needed.
I've found in the embedded space that people sell lots and lots of products that do everything you could ever want, and the most efficient thing to do is not buy those things and instead find a way to do just the subset of things you care about with your own back-end systems. The upshot of that is that because you're in total control if something goes wrong you can fix it without getting 6 people on a phone call to point fingers at each other.
As niche SaaS provider, I'm trying to avoid succumbing to the same fate. The product I built carefully for years would now be within the reach of a senior dev with a couple focused weeks -- if they knew all the requirements. To avoid being overtaken, I'm working to increase my customer's requirements -- getting them hooked on new reports and features I never had time to build before LLMs could do it for me. This makes it less likely for a competitor to be able to afford to quickly replace me.
At the same time, I have no idea what the cost of LLMs usage will be in the future. So I'm working to ensure the architecture stays clean and maintainable for humans in case this kind of tooling becomes untenable.
there's no shortage of software engineers, if it was so easy for an organization to replace a saas with something built in-house they'd be doing it all the time. In my experience in enterprise consulting implementing a well defined requirement is the easiest part. Getting everyone to agree on the requirement, getting it defined, and stopping it from changing after every demo is the hard part.
I don't think it is killing SaaS. I have definitely had to extend my sales cycle when a potential customer vibe-coded a quick fix for a pain point that might have triggered a sale a few weeks earlier, but eventually the benefit delivered by someone else caring about the software as their entire mission really wins out over a feature here and there.
If you are selling SaaS consider that a vibe-coding customer is validating your feature roadmap with their own time and sweat. It's actually a very positive signal because it demonstrates how badly that product is needed. If they could vibe code a "good enough" version of something to get themselves unstuck for a week, you should be able to iterate on those features and build something even better in short order, except deployed securely and professionally.
Everyone's going to talk about how cool their custom vibe-coded CRM is until they get stuck in a failed migration.
Yeah I have been saying this since the start of vibe coding, Saas companies rely on their sales, who are good enough to sell ther products even in tougher conditions. Software costs for the companies is 100% tax deductible, and they spend a very little on it (Most of times its less than 1% of CapEx). Only reason to optimize this cost is if the Execs of those companies think you can sell the same product.
> Everyone's going to talk about how cool their custom vibe-coded CRM is until they get stuck in a failed migration.
Failed/partial/expensive migrations is the name of the game with SaaS as well. Lock-in is the bottom line.
Migrations become much less scary when you truly own your data and can express it in any format you like. SaaS will keep sticking around, especially those that act like white-hat ransomware.
The other thing is bringing in the knowledge about what other customers in the same field want. For business-focused software this can be a boon, customers often can't really envision the solution to their problem, it's like the Henry Ford attributed "If I had asked people what they wanted, they would have said faster horses"
> What they don’t know, though, is that a poorly architected system will fail, eventually. As every senior programmer (eventually) understands, our job is complex because we have to understand the relationships in the real world, the processes involved, and the workflows needed, and representing it in a robust way to create a stable system. AI can’t do that.
I have a strong feeling the future's going to look like this:
Company vibe codes to replace a SaaS.
Little do they know this creates a time bomb: fragile systems where fundamental architectural defects are papered over by humans who knew the underlying dynamics but didn't articulate them well enough during the initial "vibe-architecture," so they're forced to patching the "impedance mismatches" with data entry or with even more vibe coding.
Those humans are eventually laid off, because of course they are. Data quality rapidly deteriorates. Operational mishaps deteriorate relationships with human counterparties. Defects begin to cost thousands to millions.
Suddenly, there's demand: not for SaaS, but for actual service businesses. Consultancies that can parachute in, do actual domain-driven design, and un-vibe that code. They do have a stronger-than-ever pool of out-of-work engineers (many from the failed SaaS companies).
The SaaS companies that survive understand that the first S no longer stands for Software; it stands for Solutions.
You're assuming that code in the future will need to be unvibed. Either the code will be good, or AI will be good enough to unvibe it. That might be awhile in the future but it will happen.
just point Claude Code at a Claude Code codebase you forgot about for a few weeks with no plan.md or agents.md or memory.md implemented
configure the session correctly this time and put it in plan mode, it will deal with your whole database schemas, migrate, tie everything to the new models correctly, make a backend deployment script, and fix your UI/UX
The panic over SaaS vs AI is simpler than people think. For years, we’ve been paying "Enterprise" prices for tools that are essentially just a UI on top of a database.
I'm a solution architect, and we recently looked at the $30/user/mo price tag for legacy test management tools. It’s insane. Why am I paying a "per user per month tax" for a glorified spreadsheet when I can pay $20 for an AI agent that can build me a custom version in a week?
So, we did exactly that. We used Claude and Cursor to "vibe-code" EZTest. A 100% open source, self hosted alternative that does 90% of what the expensive SaaS tools do, but with zero recurring fees and total data ownership.
The market is crashing because the "Application Layer" has been commoditized. If you can build and own your infrastructure for the cost of a few API calls, the era of renting basic software is over.
We aren't just building a tool; we're proving that the "SaaS Tax" is now optional.
Literally everything is a UI atop a database. Writing the data models, crafting the right uI and the right flows is non trivial. It requires iteration and that refinement is what customers pay for. The reality with SaaS stocks is they dont make anything critical. eg Take any large consumer tech company and they dont use Klaviyo. These companies build their own stack and one thats intimately connected with the data lake. It works for tech companies since they hve good data and it doesng work for non-tech companies because a good fraction of their data is trash.
> How to keep asking customers for renewal, when every customer feels they can get something better built with vibe-coded AI products?
Wrong take. You don't need to build something better, you only need something good enough that matches what you actually need. Whether you build it or not and ditch the SaaS is more of an economic calculus.
Also, this isn't much about ditching the likes of Jira not even mentioning open source jira clones exists from decades.
This is more of ditching the kind of extremely-expensive-license that traps your own company and raises the price 5/10% every year. Like industrial ERP or CRM products that also require dedicated developers anyway and you spend hundreds of thousands if not millions for them. Very common, e.g. for inventory or warehouse management.
For this kind of software, and more, it makes sense to consider in-housing, especially when building prototypes with a handful of capable developers with AI can let you experiment.
I think that in the next decade the SaaS that will survive will be the evergreen office suite/teams, because you just won't get people out of powerpoint/excel/outlook, and it's cheap enough and products for which the moat is mostly tied to bureaucratic/legal issues (e.g. payrolls) and you just can't keep up with it.
Having participated in the build of an inventory system / system of record for a large national retail company, I can't see vibe coding helping anything more than the prototyping in the discovery / requirements gathering parts of the process.
The sheer volume of data, the need for real time consistency in store locations, yada yada means that bad early decisions bite hard down the road.
Lots of drudge work can be assisted by AI, especially if you need to do things like in ingest excel sheets or spit out reports, but I would run far away from anything vibe coded as hard as possible.
One of my clients spends 500k+ on XXX licensing per year (for a 200M revenue company that's not peanuts), and on top of that has to employ 12 full time XXX developers (that command high figures just for their expertise on that software while providing very little productivity) and every single feature takes months to develop anyway. Talking about stuff like adding few fields to a csv output.
So the total cost of XXX is in the 2M/year range, and it keeps ballooning.
My (4 men) team already takes care of the entire warehouse management process except inventory, the only thing that XXX provides, we literally handle everything: picking, manufacturing, packaging, shipping phase and many others.
In any case, nobody has mentioned vibe coding.
I stated that a handful of good engineers with the aid of AI in a couple of months can provide a working prototype to evaluate. In our case it's about extending our software that already does everything, except inventory management.
When you spend 2M/year on a software (1% of your revenue), growing every year by 100/150k it makes sense to experiment building a solution in house.
Its funny you mention excel, I see vibe coding in the business sense right now being a gateway to replace all of the ad hoc uses of excel. We've basically leveled up the quality of the software you can build before buying a SaaS product or a hiring an in house engineer.
"For example, to create a data visualization I won’t seek any SaaS. I’ll just code one myself using many of the popular vibe coding tools (my team actually did that and it’s vastly more flexible than what we’d get off-the-shelf)."
That maybe doable in your 10-people startup, Namanyay. Try doing it in a larger organisation with layers upon layers of firewalls, databases, authentication systems and not the least importantly - management. Not to mention the vastly different audience, both in size and interest. Your own experience is not the experience of everyone else.
I guess they mean BI, but for a company of any scale, they aren't paying for a chart, they're paying for a permissions system, query caching, a modeling layer, scheduling, export to excel, etc.
Stand alone BI tools are going to struggle, but not because they can easily be vibe coded. It'll be because data platforms have BI built-in. Snowflake is starting down this direction and we're (https://www.definite.app/) trying to beat them to it.
The term "B2B SaaS" is doing a lot of heavy lifting here and I think conflates two different things:
(1) Business model: hosted software you pay monthly for (vs self-hosted/one-time purchase)
(2) "Glue" products: tools like Monday.com that primarily provide synergy between data sources and workflows
The article is really about (2) - and yes, those are vulnerable to vibe-coding. If your product's core value is "we connect X to Y and show you a dashboard," that's now a weekend project.
But there's a huge category of SaaS where the value is in the product itself, not the integration layer. Take Excalidraw - fits the SaaS model, but try vibe-coding a collaborative whiteboard with real-time sync, proper data persistence, conflict resolution, export formats, etc. The hard problems aren't "connect API A to API B."
Or PostHog - sure you could vibe-code some analytics tracking, but building reliable event ingestion at scale, session replay, feature flags with proper rollout controls? That's years of engineering.
The "vibecodeable" SaaS products were always somewhat commoditized - AI just accelerated the timeline. The ones solving genuinely hard technical problems seem a lot safer to me.
Simple CRUD app sure, but we're nowhere near being able to vibe code even a relatively low-complexity enterprise SaaS product.
If it's got customer data in it and/or you're making important business decisions based on it, you really need your system to be accurate and secure. My experience is the people who procure enterprise software know this and tend to care a lot about it. They often have legal and contractual obligations around that.
In the 1990s there were people who thought OOP with point and click tools like FoxPro and Delphi would make it so easy to create software that everything could be built in-house without expert programmers. The invention of SQL was supposed to eliminate roles like Report Writer and Data Analyst because now business people could just write their own queries "in English" and get back answers.
> In the 1990s there were people who thought OOP with point and click tools like FoxPro and Delphi would make it so easy to create software that everything could be built in-house without expert programmers. The invention of SQL was supposed to eliminate roles like Report Writer and Data Analyst because now business people could just write their own queries "in English" and get back answers.
And yet, precisely that happened in the end, just not with the tools envisioned. Excel, VBA and, where you had one knowledgeable employee, MS Access makes for incredibly powerful and incredibly hard to maintain "shadow IT" - and made even more difficult when someone sneaked in a password, because that takes a bit of an effort to remove [1], knowledge that is easy for us today to find, but not when I was young.
Also, back in the IE6 era, there was a lot of point-and-click created web interfaces... just that it wasn't HTML5 or even HTML. It was an <object> tag loading some ActiveX written by some intern in VB6, or Java, or Flash. I sort of miss that era but also, it was a damn security nightmare. Flash with its constant stream of security vulnerabilities was ripe for exploits, but at least it didn't run native code with full user privileges by design. I'm not kidding, theoretically you could go and import/use functions from any system DLL up to and including Kernel32. OLE/OCX, ActiveX... a design way ahead of its time.
I think opensource is a good analogue here. For many SaaS products, you don't even need to vibecode anything - there is already a reasonable OSS alternative. Yet people still pay for the SaaS. They want support, maintainability, security, edemifcation, a throat to choke, regulation and domain expertise, etc.
I do think like this HN post (https://news.ycombinator.com/item?id=46847690) is a good example of where a custom more domain specific solution makes a lot more sense that dropping in an off-the-shelf ERP. Still though, I think the bakery would prefer to buy the bakery-ERP than build it but vibecoding does reduce the barrier to entry so we might see more competition and share taking from incumbents by domain-specialized new entrants.
Lots of companies buy saas, and then spend years customizing and effectively building what they thought they bought. And for big companies, it is costly - a few tens of millions for saas licenses, and maybe around 50-100m for system integrators leaching on the enterprise, and doing the integrations and customizations, usually dancing around the data model, api surface limitations of each of the saas tools they want to wire together.
I dont think going back to having own developers, owning the code is going to be a bad financial propositions for such companies. My company is now one month into trying this out and so far, so good. We kicked our outsystems addiction and are just went live with a react rewrite - and are well into rewriting an expensive to run document management system which we were at the same time under-utilizing and abusing. Our product people are loving it since for the first time in ages we dont need to tell them "well that would be real hard, considering we have salesforce crap underneath and it just doesnt do this or that well".
My thoughts are sort of similar. AI/LLMs have allowed many of the typical SaaS customers to think that maybe they could do at least some of this work themselves, and get better results.
For most companies this was always true, but LLMs have given them the confidence to actual start writing more software in house. The SAPs of the world have nothing to fear, companies aren't going to vibe code a CRM, but they are going to be able to more easily write integrations. At a previous job we frequently had bills for €10.000 for small integrations into our ERP, but once we figured out the API and gained more confidence in our abilities, all those integrations got pulled in house.
If your SaaS platform provides actual benefits, then you don't need to worry. If your business in writing integrations for other companies into platforms you don't own, yes, AI is going to hurt your business.
This should have happened regardless of AI though. The idea that companies (over a certain size, e.g. ~20 people) doesn't have a least on developer employed, regardless of industry always seemed like a missed business opportunity. We wrote so many tools for sales, warehousing, customer service and accounting and it's hard to imagine the business functioning without those tools. I might have spend two weeks writing a tool, but if it saves sales 20 hours a week punching in orders, we get a positive return in a month.
I'd rather not name names - but one of the major, popular ones.
We contracted a lot of usage, and are using it literally like a S3 bucket with a malware scanner attached to it, and ignoring the dozens and dozens of document management capabilities it has - that we don't need. (Because really, we only ever needed a S3 bucket with a virus scanner...). This alone will allow us not to renew that contract, and save, maybe, around 2M per year.
Sure, we will have to have our own API that will require support and what not, but... we already HAD to have our own API that requires support and what not, since we have a bunch of legacy document management platforms running in various countries, and we anyway have to operate an abstraction and a router.
I am sure ours might not be the most typical case, but there will be savings, and since the economy is what it is, my bosses are telling me to go for every saving I can find, and thats one of them.
(I'd not try to re-write an ERP system, for instance, or a CRM. But a lot of smaller things where we pay a substantial premium? Sure - we will try.).
Maybe it's mostly from AI, maybe it's mostly general economic cutbacks. I also feel like these "wrapper" style SaaS products are the first ones companies are dropping when they are looking to cut costs, and I think a lot of companies are looking to cut costs. I do agree with the overall conclusion either way, that System of Record products/companies are the most likely to survive. There are a lot of SaaS companies with questionable long-term businesses who are getting hit, but that was bound to happen.
The stocks of a lot of these SaaS companies were priced on the expectation that they could become the next IBM: become entrenched with the customer and then hike prices until their eyes bleed.
A lot of companies have been too smart for that, and a lot of SaaS offerings are too small to be truly entrenched. Arguably the investment horizon is too short (IBM took decades getting to that point).
The only real vendors who managed to become the next IBM are the cloud providers.
I think it's a combination of budgeting, upward price pressure from the SaaS companies themselves, plus bringing things in-house through vibe coding, but there's another factor that I think is harming existing SaaS products. Many of them are becoming legacy solutions with AI bolted on top so they don't really feel that effective or next-level. The underlying tech might even be a generation older too - but the SaaS value-add is providing support, scaling, etc to maintain whatever some old tech that's still a requirement. At some point someone looks at all of these interconnected systems and just says 'start over'.
Vibe coding might not be supplanting all SaaS solutions but it's definitely shaking out "last-gen" solutions.
Analytical systems. I see a lot of add-on services that will add intelligence/analytics/etc and companies try them out to solve some issue they have and bounce off them frequently due to growing costs. I can only assume as mentioned that over time these are also easier for companies to in-house vibe-code as well, I just haven't seen a ton of that yet, but people are definitely trying which still shrinks the available pie.
This isn't happening. The past six months has been rough on public B2B SaaS valuations, but the impact is a lot wider than just B2B SaaS (its all non-S&P10 software), and valuations are just vibes in the end. Most of these companies are, financially, doing pretty well; seeing key metric growth, including revenue and profit. This makes sense: AI does not fundamentally change the bargain SaaS brought to the table, that companies would rather pay someone to solve their problems than solve them themselves. However, the stock market doesn't care about this. The stock market doesn't care about anything; it behaves irrationally and non-sensically, and trying to derive any sense of how stable, strong, or successful a company is from stock market valuation is like using lines of code to claim that a software project is really good.
>that companies would rather pay someone to solve their problems than solve them themselves.
Are they not able to just engage AI to solve those problems now? E.g. this morning I saw an app that did something interesting to me for $20 a month. 20 minutes in Gemini and I had a functional app that replicated the behavior. SaaS are more complex but give me a small team and a couple months and we could replicate most any of them.
This feels a lot like the old RPA hype cycle to me — more sales narrative than structural change.
Most companies are not going to replace stable SaaS with a pile of AI-generated internal tools. They don’t want the maintenance or the risk.
If there’s a real B2B game changer, it’s Microsoft.
The day Excel gets a serious, domain-aware AI that can actually model workflows, clean data, and automate logic properly, half of these “build vs buy” debates disappear. People will just solve problems where they already work.
Excel has always been the real business platform. AI will just double down on that, not kill SaaS.
> The day Excel gets a serious, domain-aware AI that can actually model workflows, clean data, and automate logic properly, half of these “build vs buy” debates disappear. People will just solve problems where they already work
Best they can do is more adware in windows. Sorry.
"The SaaS model was built on a simple premise: we build it once, you pay forever."
I've never seen a SaaS product that fits this description. There are always things to do. Libraries to upgrade, performance bottlenecks to diddle around with, an endless stream of nonsense feature requests from people at the customer who never actually use the product, fun experiments your developers want to try out, and so on.
The hard part in SaaS is to delete code, and that's what you should do, at least some of the time. Either through simplifications, or just outright erasing functionality that very few if any of your customers rely on.
What you should not do is let your customers grow the liability that is code in your production environment, unless your entire product set is designed to handle things like this, e.g. the business models of Salesforce and SAP.
It's not and I really doubt it will, for true SaaS platforms. A desktop .gif recorder (frequent example I've read about) is not a SaaS, even if you charge monthly for it.
Let's put an example an exception-tracking SaaS (Sentry, Rollbar). How do the economics of paying a few hundred bucks per month compare vs. allocating engineering resources to an in-house tracker? Think development time, infra investment, tokens, iteration, uptime, etc. And the opportunity cost of focusing on your original business instead.
One would quickly find out that the domain being replaced is far more complex and data-intensive than estimated.
There are many cases where the company might only use a fraction of the features (and therefore complexity) of the SaaS and so only need to develop and maintain those features they actually need. That's when ditching the SaaS can make sense if you can easily develop/maintain what you specifically need on your own with AI assistance.
Even if they use it less, if you combine all of the Saas products used by a company, thats a tiny fraction of the overall CapEx. And this cost is tax deductible, so there is no reason to optimise it unless Execs are really penny pinching, but at that point that company isn't worth selling to anyway.
What the authors of this kind of doomer-type articles do not realize is that B2B software companies have the data that their customers pay for, and they also have access to the AI tools themselves, meaning they can accelerate in adding new features to their products, making them more competitive.
It's a fallacy to consider the bad performance of software stocks as a definite sign that AI is going to replace them. One needs to factor financials into the equation to explain a downtrend. Take Figma for example, spending 109 mil on AWS bills, cutting through their margins. Investors know that such costs do not simply go down due to the vendor lock-in companies experience when using cloud services.
Claude Code is good, but definitely far away from being able to vibe code Figma.
I think the corrective to this is that many of these incumbents will fail to re-conceive their product stack from a user-centric perspective, and as a result they will be reduce to just a dumb data layer which is easily swappable.
Sure, they could do that, but the cultural change required is an order of magnitude harder than just sticking an agent on top of their source-of-truth and believing that the problem is solved.
Maybe it works for areas where the application is a relatively self-contained island of productivity. Figma is somewhere that a designer spends a lot of their day, so it's going to be less vulnerable, but most pieces of softare fit into broader workflows. So for Figma the disruptor is less likely to be "AI-powered designer" and more "AI-powered web builder" - e.g. Lovable or even Claude Code itself that just generates great designs.
> can accelerate in adding new features to their products
If anything AI helps companies escape the "feature wheel" that is used to justify exorbitant costs, while providing debatable (and often even negative) utility to the end user.
Keep in mind, Excel '98 would still probably be overkill for 85% of people's needs in 2025.
What companies thought was adding more utility, was actually just continually stacking costs in front of getting to that core functionality.
AI replaces the core functionality, and the "feature" scheme collapses...
“Overkill for 85% of people” is a vibes statistic, not an argument.
Also you're generalizing some things to the whole sector like every software provider now is useless and the new features they add are not bringing any values. How can you make such generic statements about a sector that is so diverse?
I start to suspect some of you are just private-equity-sponsored accounts trying to push Anthropic propaganda over the Internet.
Claude Code doesn’t need to vibe code Figma. It _is_ Figma.
Take a couple screenshots of your legacy app, write a short paragraph describing it, and the web tool will give you a self-contained HTML file that’s a fully interactive mock-up.
But it’s still a mock-up. So software dev in general will be fine, it will evolve. But unless the AI companies run out of money and it turns out the $20/mo plan actually costs $1000/mo without VC subsidies, Figma is cooked.
so many conditions need to be met to have Figma "cooked". Plus you forgot that the underlying LLMs need to get exponentially better, hard to do that without infinite VC money.
You need to have tremendous agency/will to start competing with a public company. Plus, you need to have a lot of distribution channels, competing with sales people that do this for a living since ages, and marketing budgets that are higher than your annual Claude spending.
Regular people do not press 10 clicks daily to track their calories, and you're saying that they will start competing with Salesforce and the like?
Focus is a currency and you have a limited amount of it, if all SaaS is built internally, teams would go bankrupt. There's likely always going to be a band of experts focused on solving a problem and everyone pays them to solve it for them, because they do it better and can handle the hassle of maintaining it.
AI isn't killing SaaS exactly, but instead of selling UIs, SaaS companies are going to have to focus on infrastructure and data. You have to host stuff somewhere, so there's an inescapable cost and transaction that has to take place. If businesses can pay one bill for infra + data management and get nice apps and stuff on top of that (without being locked in), that makes more sense than trying to roll stuff together even if you have a platform team.
Anecdata sample size of one, but this is not my experience at all. My business has only continued to grow over the past couple years, and I don't think I've had a single customer mention AI to me at all (over the phone or email).
It's the opposite IMHO. AI is enabling a lot more B2B SaaS. There are a lot of companies that are running on outdated software. Especially in manufacturing and industry. They've had decades of experience with very expensive IT projects where cost got out of hand because things just are very complicated in the real world.
There are many millions of companies that are going to be re-examining if they can do better in the next years. The work will still be very complicated but with the help of AI, small IT shops might just deliver enough value to be worth the trouble.
The notion of e.g. busy floor plant/logistics managers vibe coding this themselves is silly. 1) they don't have the time; these people are super busy 2) they lack the ability. 3) they'll want it done properly 4) their employers won't skip all the certifications, iso stuff, and what not.
Companies invest in SAAS software if it delivers some kind of revenue/profit benefits. If it's too expensive/complex, it can't do that. AI tools lower the cost of SAAS solutions. So the totally addressable market grows. Companies will want to maximize their ROI though. So, they'll do the usual and engage software companies and integrators to help them do this. They'll expect to pay less for more. And there will be lots of haggling around that topic. But there's an enormous amount of companies that are quite far behind on getting their operations into this century in terms of IT already. There are going to be early adopters looking for early successes here that will put pressure on their competitors if they are successful.
I'm running a small company in this space. We're seeing a lot of opportunities right now. And AI is making my work massively easier already. All those complex ERP integrations just became an order of magnitude easier to do with a small team. They are still hard though. Forget about vibe coding that. You need a plan.
The framing is a bit dramatic but the underlying shift is real. What AI actually kills is the "wrap an API in a UI" SaaS model. If the value is just presenting data nicely or doing simple transformations, an agent can replace that.
What survives: products with proprietary data, strong network effects, or deep domain expertise baked into the workflow. The moat moves from "we built a UI" to "we understand this problem better than anyone."
I run 4 side projects and the ones getting traction aren't the ones with the fanciest AI features - they're the ones solving specific problems people have repeatedly (meal planning, meeting search). The AI is the engine, not the product.
The real risk for B2B SaaS isn't that AI replaces your product - it's that your customers can now build a "good enough" internal version in a weekend with Claude Code.
Fair callout. I've been writing too many product descriptions lately and it's leaking into how I write comments. The actual point was simpler than I made it sound: AI kills "UI wrapper" SaaS, but products with real domain knowledge survive. My side projects taught me that - the ones getting users aren't the technically fancy ones, they're the ones solving boring specific problems.
It’s not as far-fetched as people think. I see so many comments here doubting you can vibe code a full CRM or e-commerce SaaS, but a skilled AI-assisted programmer absolutely can, especially if they're aware of strong open-source alternatives already out there.
For Salesforce-like CRM, there's Twenty[0], a good-enough alternative. For Shopify-style e-commerce, Medusa[1] is a headless commerce platform.
The real power comes from using AI to study how these projects implement specific features (payments, inventory, customer dashboards, etc.) and adapt them to your stack. AI excels at finding the "seams" (those connection points where a feature ties into the tech stack) and grasping the full implementation. The trick is knowing precisely where the feature lives in the code (files, functions, modules), because AIs often miss scattered pieces otherwise. That's what I'm building at opensource.builders[2]: turning OSS repos into a modular cookbook with structured "skills" that point to exact details for reliable remixing and porting.
SaaS companies are forever beholden to raising their market cap, even in solved spaces like cart, payment processing, and CRMs. Most businesses run on CRUD apps anyway, and if your core app exposes an API, you can build any customization you need on top of it. People here discounting how valuable it is for a business to have the software that runs their business on a tech stack they understand and something they truly own.
Just because it's possible to build equivalent software by vibe coding doesn't necessarily mean that companies will stop using SaaS. There are multiple reasons why...
First of all, many big companies pay a fortune to use inferior SaaS solutions instead superior Open Source solutions; possibly because one of their CTO may have received kickbacks or promises of a lucrative job at the SaaS provider as a consequence of this deal. There are a lot of politics going on behind the scenes when it comes to procurement.
Execs at big corporations are often looking for plausible ways to spend investors' money in a way that they can capture some of it for themselves. If they choose open source or they choose cheap vibe coded solutions; there is not much money changing hands. No opportunities for insiders to covertly monetize.
And then there are a lot of security implications to using a complex vibe coded app. The AI won't be able to identify the vulnerability in any decent sized codebase unless you know what you want it to look for.
Lets break this down. There is very little in newness in what Anthropic announced. Claude had skills for a long time. They have added one more layer of abstraction and called it plugins. This mainly comes with a set of integrations.
Thats the pitch.
But, what are Claude plugins?
Plugins=Commands+Skills+Integrations.
Commands are specific to Claude code. But commands and skills are nothing but prompts at their basest level.
So what is the main differentiator?
Integrations.
But what are you integrating with?
SaaS companies.
And what is the stock market doing?
Dumping SaaS stocks.
How do they think Claude cowork will work without the integrations. Without the system of records.
If anything, these SaaS products have become more important. If I was a trading guy, I would go to the github of claude plugins, see the default integrations and buy the stock of those companies.
Lot of places that I see AI disrupting - I'm not buying that SaaS is going to be a significant one.
Reading through the article:
> They were paying $30,000 to a popular tool3
Couple things we needed to understand here:
- How large is the client company
- Is that $30,000/month or day or hour....
If it's a technology company of > 1000 employees - then $30,000 month doesn't even get Finance's attention. And there is next to zero chance that anyone is going to vibe-code, deploy, support and run anything in a 1000 person+ company for $30,000 a month. SaaS wins hands down.
Any product/service that people care about comes with a pager rotation - which is 6-7 employees making > $200k/year. If you can offload that responsibility to a SaaS for < $1mmm/year - done deal.
Yeh but in a company of 100 employees for software of 30k a year, it's more than worth it to take your standard 50k (GBP) dev and have them replaced it. It's a one time cost, and the support time will certainly be less than 50% of their time every year so it saves money.
There are many companies that operate like this all over the world. Outside of the hyper-growth tech VC world cutting costs is a very real target and given how cheap Devs are outside of America it's almost always worth it.
I can't imagine it would ever be worth, under any scenario, trying to write/build/support any $25/seat SaaS software for any company I've worked at in 25+ years.
Another thing to keep in mind - very little of the cost of a SaaS license is the time it takes to build the software. Security, Support, Maintenance, Administration, backups/restores, testing/auditing said backups/restores, etc, etc.. and then x-training new SREs on how to support/manage this software, ...
Even as someone who spend 10+ hours a day churning out endless LLM applications, products, architectures from my myriad of Cursor/Codex/CC interfaces and agents - I'm dubious that LLMs will ever eat into SaaS revenue.
I'm sure (lots of) people will try - and then 1-2 years in someone will look at the pain, and just pull the ripcord.
I don't see that happening because companies need to concentrate on their differentiators. Is your enterprise vibe coding its own SaaS? Who's taking care of it?
We wouldn't do it for tools that are purpose made and have sane pricing in the market place. We do it for stuff that would traditionally go on a 'platform' like Salesforce or something that requires a lot of customization to begin with. It's so much easier to just roll your own than even just going through the procurement process of those kinds of tools much less the integration and change process (hiring consultants, etc). I'm not hands on with it, but I know our small group of AI are helping us eliminate $5m recurring annual spend this year and that's directly impacting the topic article. I won't be surprised if at some point we replace our more sticky ERP software or use this leverage to negotiate prices that are sane. Businesses have been gouged by enterprise software long enough.
Agreed. I thought it was weird on the ct ligature on something like the second sub heading and the next one had a st ligature. I stopped reading the article and was just scanning for st ligatures in the text itself (there are none) and then realised the main font was a less legible serif font and the headings that were bigger and shorter were sans serif which makes the ct and st ligatures stand out even more.
I didn't read any more of the article after that, and the primary reason was just this weird font choice.
It really isn't. Nobody's migrating their payroll vendor to some vibecoded app. I agree with Jensen Huang that this reaction doesn't make sense at all.
My own prediction is that reliable vibecoding will be additive. It's a new capability that will help high-agency people do things faster or answer questions they couldn't easily answer before. Need to spin up a big custom Monte Carlo calculation and want a simple UI to control and configure it? You can just throw that together now. Need to get a draft budget allocation for a big set of projects calibrated against a set of conflicting constraints? Let an agent crank at it for a couple hours, then review and refine manually -- or just toss it out if you don't like it.
But building, running, and maintaining production-grade services or apps that the company relies on for its basic functioning? You're not just paying the SaaS vendor for having built the product -- you're paying them to maintain it, run it, and respond to issues promptly. You're also paying them to keep building it and improving it over time. To be clear, I think there are certainly cases where the rise of "coding AGI" is going to lead companies to build some services internally versus buying from a vendor, but I predict these will be highly custom and bespoke services that are too tailored for a specific corporation to make sense for a third-party vendor to try to sell.
It seems like 'the market' is making this bet. I'm not deep into financial reports or whatever. But what I'm seeing from the tech side, this is not at all true.
If anything B2B SaaS is growing with AI, and it hasn't even begun, the biggest AI markets right now are personal. The B2B market is up for grabs for sure, 0%-1% of niches have an LLM product right now. But traditional SaaS has a huge advantage, they have reams of industry specific data, and they have the customers, sure they will have competition, but they are the incumbents.
As a founder, there is another angle here that is worth mentioning. Not only does AI B2B SaaS allow insourcing, it also allows there to be 10x (imaginary number) the number of companies building SaaS for the same use case. What we see in healthcare or finance for example is executive fatigue from demos, in many cases mostly vibe coded frontend UIs that entrepreneurs are using to test the market. This creates friction for businesses / SaaS companies that are unable to show how their solution is unique, well built or has a clear moat over the many others they have seen.
While the author is wildly overstating things, I do think AI is striking at the heart of the SaaS problem, which is the business model of "pay us $10-100+ per employee per month in perpetuity or we will hold all your data and your company's operations hostage". There is always going to be value in good software, but it is shitty vendors relying on the lock-in effect that are in danger. And good riddance.
The other issue is valuations - B2B SaaS stocks have never been rooted in reality, and the 100+ P/E ratios were always going to come down to earth at some point.
Agree on the valuations. Most have come down and many have overcorrected imho.
As expensive as some of these software seem in terms of cost per seat, most of the subscription contract rarely exceed a few hundred thousand / year if even $1mm, which is drop in a bucket for many companies. (vs running on-prem servers, having staff to support them)
You'd think Atlassian would be printing money given everybody under the sun is using them, but they only make $5B in annual revenue.
I've worked at fortune 50 companies for a while and custom enterprise software is still alive and well for things that are too business specific to buy off the product for. But they're not going to be in a rush to create their own Workday, Salesforce, Jira, Figma, SAP, etc.
Sure its fun to (vibe) code some internal version for a SaaS, but maintain it month after month? Maintain SLAs, etc? That's not fun.
Vibe coding gives you that dopamine hit of creation, but does the internal dev really want to deal with the care and feeding of the random shitty timesheet app they created?
Do they want to take on the work of integrating random backend systems that timesheet system needs to talk to? Do they want to get called at 3AM when it's down?
Even AI assisted, living year after year with production systems is hard.
> we have to understand the relationships in the real world, the processes involved, and the workflows needed, and representing it in a robust way to create a stable system. AI can’t do that.
That is because AI is living in our world, instead of the opposite where we live in AI's world.
Case in point: maybe the AI hallucinated a class method that never existed in our world yet, but perhaps in the AI led processes and workflows it would be written to better fit into the smooth gradient decent those same top parameters' scores.
I think there may be other factors killing SaaS, particularly data sovereignty.
"According to IDC’s Future Enterprise Resiliency and Spending Survey from June 2025, 45% of all organizations and 56% of “digital natives” cited data sovereignty and potential cloud changes as their greatest concern for 2026."
“Killing” is a bit strong, but is there a world where folks just vibe code solutions that they would have bought previously? Absolutely and and I think that world is here now.
I’ve seen many startups recently were it was like “guys I could vibe code your ‘product’ in the afternoon.” Yes someone needs to look after it etc, but the bar on where companies buy vs build is getting much, much higher.
(Insert rant from dev teams about the code sucks, who will maintain it, etc). Yes all valid points, but things are changing regardless of if folks like it or not.
Counterpoint, all your customer's data just got stolen by teenagers in the 3rd world and is available for purchase on the dark web. Was it worth saving $5k a month?
A lot of startups/small businesses are like "with AI we can build more than ever". The problem is so can everyone else and capitalism rewards scarcity not value. The bar for startups and small software business has risen quite substantially. I know we are avoiding buying software now where I work if possible unless we previously committed to it (contracts).
Saas companies will survive for the same reason they do today. The operational overhead of any sufficiently complicated piece of software is too much, even more so if it's vibe coded.
One problem with centrally produced and distributed software is that a small subset of users demanding certain features results in feature bloat for everyone. Costs for all features are shared by all users.
Probably one way SaaS companies will adapt is to break up their offerings into more modular low cost components. While many customers will end up paying less, the addressable market will probably increase because of the new low cost options.
To a degree but most enterprise focused software usually has differential pricing. Often that pricing isn't public so different companies get different quotes.
The framing of 'vibe coding replaces SaaS' misses the more interesting shift: the value SaaS provided was never really the software — it was workflow automation. Software was just the best delivery mechanism we had.
What's changing is that agents + APIs are becoming a better delivery mechanism for many workflows than a UI you manually operate. A company paying $50k/year for a marketing analytics dashboard doesn't actually want a dashboard — they want answers about what's working. An LLM with API access to their data sources often delivers that faster than navigating someone else's opinionated interface.
The SaaS most at risk isn't infrastructure (Stripe, Twilio) or systems of record (Salesforce, Workday). It's the 'pretty UI on top of data you already own' tier — analytics, reporting, simple automation, basic CRM. That's where the compression happens. The products that survive will be the ones that become the system of record, or that offer value AI genuinely can't replicate (regulatory compliance, deep integrations with legacy systems, etc).
I didn't realize B2B SaaS products were in freefall like his numbers suggest. I'm not convinced customers are leaving to vibe code their own products but I do believe we're seeing a major shift in the market, pushed by the sudden relative ease of coding. There are a lot of B2B SaaS products which are outdated and I wouldn't be surprised if they're supplanted by much faster competition
Yup it's definitely not because _customers_ are coding solutions, but the trend and motivation seems to come from the fact that customers are realizing there's something else possible except being tied into expensive recurring yearly subscriptions.
I was surprised when I saw the numbers from Bloomberg myself as well!
The keyword in "Software as a Service" is not Software; it's the Service.
In the early days, the tagline for Salesforce is "No Software". It's secret recipe is this: your sales team only need a browser and a credit card, to get the service. No software installation needed. Even if you have a genius can code something equivalent, it will never be a "service". That genius is not going to support it, not going to add storage for you, not going to restore an accidentally deleted record for you. That takes an army to deliver. It is a service.
Of course, Marc Benioff kind of shot himself in the foot by trying to get ahead of the AI curve... and gutted their customer service division. If the service is delivered by AI agents, what is the selling point again over other AI agents? They have debased their key strengths and are getting punished for it.
I used to be a big advocate for Salesforce in my organization. And it was really great .. allowing us to deliver new functionality without the usual IT procurement bureaucracy.
Now with cloud maturity and Vibe coders who will get better and cheaper, I think it's possible to replace all the features we use on Salesforce at a fraction of the cost of our Salesforce licensing cost.
There is no evidence presented that internally "vibe coded" products are the reason hubspot et al are struggling right now. If anything, the fact that the divergence from the broader index starts in April of last year (well before the current vibe coding moment got going) is evidence that this is something else.
Vibe coding seems to be the iPhone camera to DSLR moment for programming.
- No professional used an iPhone for years. Most don’t today.
- Professional scoffed at it as a toy
- The toy shifted the balance of volume through everyday enablement of amateurs to a degree that professional were right, but now in a severely lopsided terrain.
The value ends up in the most engaged paradigm, rather than the most perfect one.
Most people who've been in a business SaaS environment know that writing the software is relatively the easy part aside from in very difficult technical domains. The sales cycle + renewals and solution engineering for businesses is the majority of the work, and that's going nowhere.
I feel the same way about needing to build platforms that enable companies to build on top. That's why i'm building adventureflow.ai to enable energy companies to build their agentic workflows directly on top of my platform.
The expectations have completely changed now, customers expect more software for less, because it is 5-10x cheaper now or faster to build and there's no changing that. Tools like Opus 4.5 have completely disrupted the software marketplace!
Salesforces market cap today is 180,35B. That, not the change, is the market's judgement of how valuable salesforce is, despite any risk from AI. If salesforce stock drops with a new AI model that doesn't mean that investors think salesforce is worthless. It means a slight risk to their business is getting priced in.
I am kind of hoping that AI will kill the startup grab for money TBH. Too many wannabe CEOs I've met in the past 2 decades have gotten rich thanks to a lucky pitch without a clear path to a viable product. At least 6 of those I know did so at the expense of developers that accepted equity over cash...and the developers wasted a ton of time and 2 even were briefly homeless as a result...and none of them live in California.
Hopefully wannabe senior leadership will try and take advantage of AI without taking advantage of developers, because most of us just want to write code and build something great.
Reminds me of the story of when the Surgeon General (in the US) reported that smoking causes cancer.
People stopped smoking immediately, and cigarette sales tanked. The cigarette companies laughed (with all the phlegm in their throats and lungs) and sales came back 1-2 weeks later.
I suspect in a few months or a year companies with vibe-coded replacements for SaS products will find they need to go back: But, just like how many less people smoke today than in the past, the writing is clearly on the wall. At some point someone will figure out how to replace SaS with AI; it's just going to take a lot longer than many think.
I was reading through the article and waiting for the key info to drop, but nope. It never did. It seemed like marketing fluff. If anything, vibe coding may eliminate some of the B2C SaaSes, but not B2B. If you think an enterprise is going to vibe code a B2B offering that they pay millions for, you're out of your mind.
Here's my general mantra regarding AI: NEVER take suggestions about AI from people who have a vested interest in it. CEOs of companies that train and offer LLMs, Authors of Books about LLMs and AI in general, etc.
This may come off as an unpopular opinion, but this is how I felt after listening to Steve Yegge recently. He has a new book about Vibe coding and he goes on in the interview/podcast to say that the best programmers he knows in the world (the ones better than him and maybe even the top world class programmers), would be equivalent to those of interns in an year, if they don't start vibe coding or use AI. I respect the guy, but damn, this is just peak delusion. He didn't even say it as a hyperbole, he meant it.
According to popular CEOs of companies training LLMs, 2024 was supposed to be the year that would eliminate the need for Junior and mid-level engineers. 2025 happened. Now, we are in 2026.
So yeah, I'm never taking advice about AI from these people ever again.
> Here's my general mantra regarding AI: NEVER take suggestions about AI from people who have a vested interest in it
I get where you're coming from, but let's say you're talking to a HVAC installer, and he recommends you a system to get - I'm sure there's financial self-interest on his part, but I do like to think that he knows quite a bit about what he does, and believes what he's selling is genuinely good stuff (and has reason to), even if he oversells it a bit.
The analogy can work if you're not looking for an HVAC at all and the HVAC guy is instead approaching you, unprompted, to explain that you need to buy this new system. Because if you don't, your business will become uncompetitive and fail.
True. That's the case with almost any commodity in life. That's why I was specific about AI.
The difference is, in other sectors, there's no fear-mongering. If you don't use their HVAC, it's fine. Your job isn't getting replaced. The air you breathe in your home isn't going to be fully polluted. You have other options.
With AI though, there's no middle ground. You either use their tool and become extremely successful (so much that you don't know what to do with that much success) or you're out of a job and become obsolete in like the next 3 seconds.
Are there real documented cases of a company replacing their SaS with a vibe-coded version?
Like I can see how a very small company could replace a portion of an overkill and underutilized SaS platform.
I don’t see how a larger more complex business could replace their SAP or ADP with a vibecoded version.
These stories are all very similar in where the author knows some CEO of an obscure company who told them they had an engineer reverse engineer and vibecode some obscure SaS solution and saved them $50K.
With a new agentic-lashup tearing across the internet every week, pointing the way to "gradient descent" software development, any purchasing manager worth their salt is going to ask some serious questions about their enormous SaaS bill before committing to another expensive long term contract. It follows that valuations must decline. Even if only because risks to moats have increased, but also because it makes sense to negotiate hard on pricing when there's fear in your counterparty.
Preposterous. Have you ever worked for a company as a programmer or for that matter as a manager? They don't just replace products ad hoc. There's an enormous amount of due diligence that goes into any new software product - making sure it fits the company, that it's secure, that it works properly... I recently worked at a small startup who spent on sales force @ $100,000 a year. I said you know what we could do this ourselves and you know what they said as every company I've ever said that to says? "We don't want to support it. We need to cover our ass. Everybody knows how to use this"
I'm not one to sell generative AI short at this point, but this seems at odds with the longer-term trend towards more centralized, off-the-shelf solutions, like Shopify, Salesforce, Squarespace, over the "bespoke," in-house alternatives that many companies developed in previous decades. Perhaps AI is making the TCO of "rolling your own" so low that it makes sense to go back to that over SaaS?
AI, as do most things, help the big players get bigger. If someone is automating small parts of the b2b layer they get dropped, but it’s harder to drop an automation that companies are used to. I don’t see how AI is changing that, companies spent a lot of time and money to set up the automation because it’s needed and because they can write a potential replacement cheaply doesn’t mean they are going to rip away something that works and is reliable.
I would assume one major thing here is that many orgs only need a small subset of functionality from what most products provide. Many times, that small subset of functionality is only "good enough" in and of itself, but the org is paying the premium for the entire suite of whatever it is. This makes realizing that an LLM can get them to MVP and beyond much easier.
Charging hundreds of thousands if not millions per year for very basic functionality is what is "killing" b2b SaaS.
There is also the benefit of being able to use a single database (and hence schema) across multiple "apps". In many cases the complexity arises from the fact that all these apps have their own databases.
I predict the fallout from this is companies being nickel & dimed to death by a million smaller subscriptions (rather than just cutting a huge check to Workday, ServiceNow, SAP, Oracle or similar). There is such a glut of AI ISV startups that are filling highly specific niches, and some are quite good, but they're all usually in the $10-50/mo/user. Gets to be big numbers pretty quickly in a large enterprise.
There will be an exiting explosion of internal apps and tools and then it will die down as companies get back to business. All those tools and apps will turn into legacy apps and they will eventually try to migrate data off them into some saas platform because its not core to business.
Saas will have to drop prices to be competitive so fat margins will go away which will probably affect margins for AWS etc.
Sure, vibe coding has impacted user's expectations. They know you can ship a new update easier and faster than before - and you actually can.
But, not sure which successful SaaS companies just stopped shipping any updates to the product, never talked to their customers and never added any new features to win over major new accounts - and still managed to survive and thrive?
And the author actually confirms this:
> AI isn’t killing B2B SaaS. It’s killing B2B SaaS that refuses to evolve.
It's this generation's "build vs buy." I imagine it will play out the same way, like a revolving door. Customers churn because they can "build it themselves," then a year later when they're sick of maintaining a mess of code for some internal system instead of delivering value to their own customers, they come back. A blip.
Or maybe there is look at true value proposition of the B2B SaaS products. It is not big spend per user, but it does add up eventually. And then savings start to look like big numbers. Big enough for some manager to justify action. This might lead cutting some seats just on cost basis.
- A company vibe codes their own app to replace a SaaS. Great when they only wanted a small chunk of the functionality.
- Startups benefitting from AI coding are copying mature SaaS companies and competing on price.
- Mature SaaS companies are branching out into each others domains. Notion is doing email. Canva is doing an office suite.
here's the secret saas can vibe code features too on top of their paid well developed and secured api. they can get off their ass and vibe code a mcp wrapper, so user can use the ai tooling they pay for to interact with their saas. and they'd be called visionary hero of the agentic revolution.
but they don't want to. and they will be replaced, as it's good and well.
Maybe the type of SaaS that's akin to stock media (photos, video, music). Just hard enough to do from scratch, but not important enough that it needs to be exceptional in it's field. I've made some money off software like that, and it was nice, but I always knew it couldn't last. Better developers took most of it from me years ago.
there is no saas downturn caused by AI. wall street is just starting to say hang on a minute, why is this SaaS stock trading at a price to earnings ratio of 300?
then the sell-off is attributed to AI because it is far easier to say to shareholders hey we know our company lost half its value but thats actually a good thing because we need to pivot to AI and we're going to spend all our free cash flow on AI software and our stock should totally be trading at 300x earnings again in a few weeks. if you can last another few months as CEO and the fed cuts rates you'll be able to ride it out
of course, the tide is going out on a few dogs. I don't think adobe will become dominant again
you see the same trend with mass-layoffs being blamed on AI. easy way to sell bad news to the shareholders
in 2026, AI and JE are the two reasons for absolutely everything
Something notable for SaaS which this article doesn't mention is that in some cases the reason to buy rather than make yourself is due to needing to handle a bunch of different regulations which LLMs don't threaten (barring businesses which would rather have lawsuits than pay for a SaaS).
We will have to go through the stage of disillusionment of what AI is and understanding of what it is not. There is too much FOMO and too many stories driven by the heavy-AI-invested parties today to see thing clearly.
The real story isn't that some enterprise mega corp is going to vibe-code their own Workday.
The real story is that SOME startups made up of one person (or small number of engineers) will do it, and create pricing pressures against Workday, or DocuSign, or other B2B SaaS.
But if you are a mega corp, spending 0.1% of your CapEx on SaaS subscriptions, do you really want to switch to products made by some no name one person startup? You might save 0.05% of your CapEx, but if something didn’t workout, your whole business will be screwed.
To your point, I think there are 3 main categories:
1. Too big to fail (SalesForce, ServiceNow, ServiceTitan, Shopify). They’re targeting megacorps’ core business operations. Switching costs are too high. They will survive,
2. B2B non-core (PagerDuty, Vanta, Monday, Atlassian). They’re going to have stiff competition and most here will fail or merge/consolidate. They have the most to lose because they’re non-core to a business’s success and pricing pressures will cause many of them will be easily vibecoded with enough time. The large TAM here will attract hundreds of competitors each.
3. Consumer SaaS (Notion, Canva, Grammarly, Dribbble). They're good as dead and can be buried.
A middle 100-500 heads firm don't need enterprise level SaaS, a vibe coded website will suit them better.
Fundamentally, those workflow/orchestration SaaS needs to answer the question why people should pay you premium while only getting 80% where they want to be.
For the most part, you can replicate any B2B SaaS product in a spreadsheet. The same reasons why spreadsheets didn't kill B2B SaaS apply to "in house vibe coded SaaS replacements." The original in house apps are (and continue to be) spreadsheets.
It kills low hanging fruit SaaS that was always just some minor piece of functionality wrapped in a subscription model usually based on 99% open source and 1% actual contribution.
I think people here need to accept that software is becoming electricity, you get charged when you use it and by how much. You don't pay for a box shaped electricity or purple color electricity, it is just electricity.
Sounds like a lot of products will be in trouble if AI becomes more advanced with producing secure code with maintenance processes to keep evaluating code on an ongoing basis
How corporate incentives are aligned will also define the trajectory. The person who is going to take the call whether to go for vibe coded tool or external vendor has to have enough incentive to put this neck on the line.
Imagine this, you are VP of finance. You know you can get a nice tax calculator built quickly using vibe coding, but will you put your neck on the line and say, let's replace the existing vendor and use my vibe coded tool. You might fired if you send a wrong tax report to your auditor.
Let's take another example - VP of sales wants performance report and visuazliation of the sales team. He has 200 BDRs. The daily sales standup depends on this report, team gets yelled at or praised basis this. Now, will this VP be willing to put his neck on the line and say - let's use my vibe coded report and discard the existing SaaS. Even if such a report feels trivial, it is vital for functioning the sales team and hence, it needs to be reliable.
I think vibe coding will work at prototype level - the trust gap is very huge right now to even consider it for internal tools.
And, until vibe coded tools are stress tested enough this trust gap will not be fulfilled. Think about this, why some of the biggest companies in the world still run on softwares built in 2000s.
The reason for divergence is actually much simpler. NASDAQ 100 includes data center builders, Morgan Stanley software index doesn't. Stock market is going down across the board if you exclude data center construction.
Remember when businesses ran on cobbled together access databases and vb? It was easier than building something ny prompting an llm.I made a good living just rewriting those things for them when they fell apart.
This was the same hype around NoCode a few years ago. Thing is people for software as a service and internal teams dont want to do the service part. A common thread among all of the vibe coded stuff I hear is that nobody talks about the next requirement which was handled. Frankly UIs are learnable and after a few tries everybody gets used to it. The extra polish is totally not worth it. Eventually every SaaS on the non-critical business path gets ripped out with one of the platform players such as SAP, SalesForce etc which are infinitely customizable. Salesforce even has its own language in which customers write their apps.
The system of record bit is quite correct and imo the only durable advantage. But that doesnt explain why Klaviyo shoyld worry. In fact marketing workflows are some of the most important systems of record directly tied to business earnings.
The assumtpion that in SaaS you build once and everybody uses it is false. There are always customizations or features that need to be built for big contracts. Its the small players that are okay with out of the box SaaS since they lack the budget to pay for a customization. Ironically, these teams will get hurt the most should they choose to go down the vibe code everything path. Here infra capex is far lower than payroll and owning too much software (vibe coded or not) will simply steal time from business development.
Honestly, the market just panicked and caused a temporary correction thats all. Then we all got busy and wrote a bunch of thought leadership crap on top.
In reality AI is the best thing to happen to SaaS companies since it reduces RnD time and expenses, allows even small players to bid for large contracts without overloading existing dev capacity and also makes it possible to put devs in front of customers since now there's no need to sit down and code. AI is also the best thing to happen to engineering since it now allows for product management to be folded into the existing craft.
SaaS’s competitive advantage is not tech but distribution.
Just because the stock market dips, and a bunch of media outlets ascribe this to AI doesn’t mean that AI ‘caused’ the dip.
If people really think non-tech businesses will start running their own software organizations and maintain all these tools themselves is an idiotic assumption at best.
The “building” of software has never been and will never be the bottleneck.
I am building a desktop app that deploys and manages high availability postgres instances on VPS right now..
The AI is constantly failing at helping me because its about generating and managing remote configuration that is very fragile if something is missing.
I don't think AI is killing B2B SaaS so much as companies are finally reckoning with the immense costs of SaaS in a markably different environment than when SaaS exploded in popularity, and AI offers an off-ramp to some. Let's break it down camp-by-camp to show you what I mean:
1) The must-haves. These are your email and communication systems, the things you absolutely have to have up and available at all times to do business. While previously self-hosted (Exchange/Sendmail, IRC/Skype/Jabber, CallManager/UCS), the immense costs and complexities of managing systems ultimately built on archaic, monolithic, and otherwise difficult-to-scale technologies meant that SaaS made sense from a cost and a technical perspective. Let's face it, the fact nobody really hosts their own e-mail anymore in favor of Proton/Microsoft/Google/et al shows that self-hosting is the exception here, not the norm - and they're not going anywhere regardless of how bad the economy gets. These are the "housing stock" of business, and there's plenty of cheap stock always available to setup shop in without the need for technical talent.
2) The juggernauts. The, "we can do this ourselves, but the pain will be so immense that we really don't want to". This is the area where early SaaS solutions cornered and exploded in growth (O365, ServiceNow, Google Workspaces), because managing these things yourself - while feasible, even preferable - was just too cheap to pass up having someone else wrangle on your behalf with a reasonable SLA, freeing up your tech talent for all the other stuff. The problem is that once-focused products have become huge behemoths of complex features that most customers neither need nor use on a regular basis, at least after the initial pricey integration. Add in the ease of maintainability and scalability brought by containers or microservices, along with the availability and reliability of public cloud infrastructure, and suddenly there's more businesses re-evaluating their relationships with these products in the face of ever-rising prices. With AI tooling making data exfiltration and integration easier than ever from these sorts of products, I expect businesses to start consolidating into a single source of truth instead of using dozens of specific product suites - but not toppling any outright.
3) The nice-to-haves. The Figmas, the HubSpots, the myriad of niche-function-high-cost SaaS companies out there making up the bulk of the market. Those whose products lack self-hosted alternatives risk having vibe-coded alternatives be "good enough" for an Enterprise looking to slash costs without regard to long-term support or quality; those who compete with self-hosted alternatives are almost certainly cooked, to varying degrees. If AI tooling can crank out content similar in quality to Figma and the company has tech talent to refine it for long-term use, why bother paying for Figma? If AI tooling can crank out a CRUD UI for users that just executes standard REST API calls behind the scenes, then why bother paying for fancy frontends? While it's technically interesting and novel at how these startups solved issues around scaling, or databases, or tenancy, the reality is that a lot of these niche products or services could be handled in-house with a container manager, a Postgres instance, and a mid-level IT person to poke it when things go pear-shaped. The higher per-seat prices of a lot of these services make them ripe for replacement in businesses comfortable with leveraging AI for building solutions, and I expect that number to grow as the tools become more widely available and IT-friendly in terms of security.
Ultimately, the core promise of SaaS to business customers was all the functionality with none of the costs of self-hosting support. Nowadays, many of them have evolved into solutions that are more expensive than self-hosted options, and businesses that have shifted IT into public clouds or container-based systems have realized they can do the same thing for less themselves, at the cost of some UI/UX niceties in the process. Now that we (IT) can crank out integrations with local LLMs with little to no cost, we're finally able to merge datasets into singular pools or services - and I'm not talking about Snowflake or its "big data" ilk so much as just finally getting everything into Salesforce or ServiceNow without having to bring in consultants.
The must-haves and many of the juggernauts will remain - for now. It's the niche players that need to watch their moats.
That's a great point. The restaurant business is one of the worst you could possibly go into, with razor thin margins and astronomical failure rates. If those are the future economics of SaaS (accustomed to 70-80% gross margins), than you ought to cash out now while you still can.
People have been debating what "real" development is for as long as there have been computers. "That's not real software development, you're not even controlling which registers you use!" "How can you say that's real code when you aren't even managing your own memory!" "That doesn't look like C, must not be code!" and on and on.
At the level of the old adage about whether the horse-drawn buggy-makers are in the buggy business or the transportation business, it's all telling the computer what to do in the context of providing a customized tool that you or others might use. So in this context, a customized Excel spreadsheet counts. And so does a vibe-coded app.
And of course, wringing our hands about what it looks like now totally ignores the fact that it's not going to be like this for more than a year or two at most.
How long until a user can reasonably say to Claude or similar, "I need Bob here to track production at my factory. Lay out a set of tools to do that, and make sure they're layered with help and tutorials so Bob can learn on the job because he doesn't know anything. Don't let him make any mistakes."
That's probably not coming next year, but certainly it's not ten years away.
Having worked in enterprise B2B SaaS for a long time, almost every feature I built could have been a simple spreadsheet or some emails. So I'm highly skeptical AI is going to change anything.
Enterprise sales basically works like this: A non-technical sales team aggressively promises everything to win a deal to a non-technical procurement or exec team. When the deal is won, the SaaS sales team tells engineers "go build this" regardless of how stupid it is. And the customer tells their employees "you now have to use this SaaS" regardless of whether it makes sense.
AI is killing this. Millennials are killing that. Gen Z certainly must be killing another thing. And Yarn? It was killing NPM. This nonsense type of title really was made for a world where nuance doesn’t exist.
B2B SaaS is projected to grow over the next decade, not shrink. Just use the LLMs and you’ll be reminded of the value provided by the company selling non-core but important tools for your business.
no. High interest rates and a cautionary view of future economic growth are killing B2B SaaS. Money is no longer free, and so there is a bigger push for cost-cutting rather than growing your buisness with free money.
A link shortener is such an easy thing to code, it's essentially one database table with a redirect. To add to that, there are many open source libraries to implement link shortening, including analytics and stuff. Even then Bitly and Rebrandly have customers (from their website) like Toyota, Cisco, Oracle, Monday.com, New York Times, etc.
Are these companies unable to build a link shortener? It's also so easy to migrate off shortener service. If they can and still choose to use these shortening services, there must be other reason. And that reason is that they simply don't want to. This has nothing to do with AI.
I run a software company and one of the reasons customers say they want to migrate from their homegrown spreadsheet is because the guy who built it left. A freaking spreadsheet!
Such blog posts and probably many comments here are the perfect answer to "Tell me you don't run a real business without telling me you don't run a real business"
Regarding your last comment, majority of the people here are costal with FAT paychecks slinging code for VCs. It’s a totally different universe than running a Saas. That said, still a valuable forum.
Until Claude Code comes with indemnity insurance for HIPAA / GDPR / etc… B2B SaaS is here to stay. You want me to convince my auditor that the vibe-coded in house software handles PII correctly?
Making the audit someone else’s problem is 90% of the ‘buy’ value in ‘build vs buy’
Spot on. You could argue that most companies buying B2B SaaS could almost always build a clone internally but they need someone to assume SLA and liability.
I wouldn't say that AI is killing B2B SaaS, let's analyze the situation: AI is shifting the boundaries of the Build VS Buy decision.
If software is cheaper... building it seems a much better option but...
...the fact that software is cheaper means that SaaS companies will be able to be more competitive with price, offering more features, integrating them much better with the newest agentic tools and frameworks that are getting released.
Sure, the market will adjust, some will be pushed away, others will find businesses that weren't possible before... but we're far away (I hope) from the AGI revolution, don't forget to see this crysis as an opportunity too.
I've worked in SaaS for most of my career, only recently working at a big corp who is largely the buyer and user of SaaS tools to meet their objectives. From the perspective of the corp business buyer, they want something that works for their needs and they want to buy something instead of build it because the support costs are gnarly. They already have engineers dedicated to the tools they've purchased. Much better to put the risk on someone else they can yell at. And the permissions and access to these tools, reports, data, is usually its own special problem to manage. Building a lot of one-off tools is going to just give IT a huge headache and they will push the org to buy before vibe coding a solution.
Look, software is not going away even though everybody thinks they're a developer now. Do you think companies are going to replace Microsoft Windows, of which there are a billion installs, or Salesforce, Crowdstrike, Quickbooks etc, with some Vibe coded AI slop an intern "coded"?
When enterprises/businesses in general upgrade any software in the company it takes years sometimes... There is also Vendor lock in, you can't just swap things with vibe-coded slop, and trust me your manager will never want to do that either because his butt will be on the line.
I disliked how SaaS CEO's were decrying the death of engineers. Their coordinated layoffs over the past years or so was excruciating to watch and experience. Their language was aggressive and inflammatory.
Although the article may also be hyperbolic, I'm not going to comment on reasons why it might be. Instead, I will agree, and think SaaS companies stock performance this year will be proof. Sure, it might not be the collapse that AI doomers are hoping for, but all the FUD they spread over the past few months to years will signal that they're not insulated from it. They made their cake, now they have to eat it too.
At the end of the day you do not need to replace your B2B SaaS with AI.
You need your B2B SaaS to think you can use AI to replace it though, so said SaaS will keep it's prices reasonable. Otherwise they have you by the balls and will charge you much as humanly possible.
If that would be true, expect in the next decade a frantic search for seclusive grey beards, those who haven't given up their rituals and ancient languages.
If your workforce is vibing all day, they will have no capacity for maintenance, because it isn't their code. So the maintenance that happens will be slop and more spaghetti. I am not saying cases like that never existed before, but such companies will face a moment of truth sooner or later.
The survival advice here — 'be a system of record,' 'make switching painful with compliance' — is exactly the playbook that makes customers want to leave in the first place. The deeper question is whether vibe-coded replacements will rot fast enough to send people running back, or whether AI gets good enough to maintain them too.
No it isn’t. Writing the code was never the issue with making software, it was designing it.
You can shit out an app with AI, just like you could with Indian workers. But that doesn’t mean it will work properly or that you’ll be able to maintain it.
And most importantly, it only works for code they could steal from GitHub. It has no idea how to replicate sensitive systems which aren’t publically documented, and those are some of the most valuable contracts.
> how they’re potentially losing an $X00,000 account just because the customer can’t use a specific failure reporting workflow in the SaaS. They’re now working with me to build what the customer needs and retain them.
You mean the poor SaaS company actually now has to implement features needed by customers??
Just a single data point, and I‘m not pretending it’s a conclusive one (yet), but I think there is a middle ground between buying a SaaS and vibe coding a replacement: replacing a SaaS with your own solution, using AI coding agents — while actually knowing what you do because creating robust software in-house is already a core competency. No vibe coding needed.
At my company, we build software every day because our business is running a job board.
We always had kind of an impedance mismatch when it comes to creating content pages (think landing pages for marketing).
Yes, we can do this ourselves, but then software engineering resources are in conflict between shipping the next feature and creating landing pages.
So we introduced Webflow. Now marketing could create their content themselves. Did it match our corporate identity? Hopefully. Was it technically sound? Sometimes. Was it fun for software engineers to fix things in Webflow. No.
It kind of worked, but wasn’t exactly ideal.
Meanwhile, software engineering became more and more productive with the advent of AI coding agents, Cursor in our case.
So we tried another approach: giving our content creators Cursor.
But that was brittle, too: Cursor presents an overwhelming complex UI for someone who never used an IDE before, it could do a thousand things while only three are actually needed in this use case, you have to explain git on top of Cursor. It kind of worked, but only kind of.
It’s like a hyper-focused „Cursor light“ in the browser, so our content creators can just „chat away“ and create their content pages, with all the guardrails baked into the product. Getting tracking pixel integration etc. right just works. Matching our style guide perfectly just works.
And as a bonus, there is a whole git based storage and versioning workflow underneath, so software engineering can support and test and deploy the results with their preferred tools and methods, but none of this complexity leaks through to the content creators.
We built this ourselves in days, not months, thanks to Cursor, but it’s not vibe-coded — it’s a rock solid application that we will support and improve long-term without headaches: https://github.com/dx-tooling/sitebuilder-webapp
No.. just.. no.
This will be a thing for like 1, maybe 2 years, then people will realise it doesn’t make sense to spend $50K of time per annum to replace a $500/month subscription for a better product.
One point is now you dont have to pay money to 3rd world countries like mine :(, you can manage with fewer resources inhouse. AI will take care of slop work.
for profit software is pretty gross tho. it can be made indefinitely complex, and people can make and sell tools to patchwork the complexity together, but then u need more tools to patchwork all the tools together ...
and u end up in aggregators aggregated aggregators type situation where optimal solutions never arise because we don't actually cooperate enough to produce them
ai is fitting into the notion that this is all bullshit ... even if not in the correct way
Honestly, I'm surprised by how people are pushing back against this idea. I feel like vibe coding is just in its earliest moments of actual viability, and my mind is totally blown by it, and it strikes me that it's exactly what I've always wished software could be. Plastic, flexible, personalized, effective, responsive, organic.
As an anecdote, I've been vibe coding an accounting system that perfectly matches with my own expectations of accounting software, i.e., it's intimately connected to CSVs, imports and exports from CSVs, but acts as a kind of enrichment and reporting and file association layer. If there was anything like this, any kind of SaaS that I could have and download as software and run on my own computer offline and be able to inspect and trust and version control so they wouldn't add or remove some kind of feature that I wanted down the line, then I would have gone with that.
But now I have essentially my absolute ideal solution, written with a Python backend and Vue and JavaScript frontend, and it's radically improved my ability to maintain accounting for our business account.
And I think there's something really important to point out here, which is that the feeling of lock-in is very seriously reduced when you are Vibe coding your own software, because if you don't like the way that it works or you realize that there's something missing, you can add it pretty painlessly. Like, that's always been a huge challenge with choosing vendors for a SaaS platform, is you think, oh no, well, what if their conceptual model for what I'm trying to do doesn't quite map onto our own internal systems or understanding of what's being done? Well, when you have your own Vibe-coded SaaS, you can just add that information. So there's something incredibly organic about it. I used to work at a startup in Redmond where we built this large internal system to manage a scientific process with lots of machinery and data, and it was incredibly empowering and actually became one of the core values of the business that was able to be licensed to other businesses in the same industry. And it seems like we're just improving that capacity from here.
Now obviously, if Vibe Coding magically were to go away or became much more expensive, then I'd have this legacy piece of software, which I couldn't improve, and that would be a dead end. But my expectation is that the functionality that we have today will only improve. And in several years, the scope of changes that I'll be able to make, the level of professionalism, modularization, maintainability, code quality, will only improve. And so this has me thinking in general that software is kind of undergoing a step change where we're moving into the so-called age of intent beyond the age of the interface. And that's tremendously exciting to me, and I just couldn't be more stoked about it.
It's a tale as old as time that developers, particularly junior developers, are convinced they could "slap together something in one weekend" that would replace expensive SAAS software and "just do the parts of it we actually use". Unfortunately, the same arguments against those devs regular-coding a bespoke replacement apply to them vibe-coding a bespoke replacement: management simply doesn't want to be responsible for it. I didn't understand it before I was in management either, but now that I'm in management I 100% get it.
OTOH, I was hired by an enterprise that was many months into a giant backend rewrite. After wrapping my head around the many plans, I realized they were rewriting Django, badly. One weekend I prototyped the whole thing… in Django. It worked. It met the specs. It was a CRUD app with a REST API.
I came in to work Monday morning, showed it off, and inadvertently triggered a firestorm. Later my boss told me not to do that again because it caused havoc with schedules and such.
So I quit and found a better job. Sometimes the new guy can make a better version themselves over the weekend, not because they’re a supergenius, but because they’re not hampered by 47 teams all trying to get their stamp on the project.
(In before “prime example of overconfidence!”: feel free to doubt. It was a CRUD app with a handful of models on a PostgreSQL backend. They were writing a new Python web framework to serve it, complete with their own ORM and forms library and validation library. Not because the existing ones wouldn’t work, mind you, but more out of not realizing that all these problems were already sufficiently solved for their requirements.)
In the beginning of my career, I've been once told by my senior management that I should never again:
1. Optimize things so that they work 10 000 times faster because it makes us look incompetent (must be done slowly to show gradual progress).
2. Brag about such optimization (to stakeholders) without first synchronizing this with them (so they can brag proportionally to their pay rate :) ).
31 replies →
I can relate to this so much. When I was a newly joined Google consultant at a partner firm, we went to their office - some 13 different types of cuisines, different types of game rooms, lounges and what not. A luxury star hotel experience. We were waiting for our meeting on behalf of this one particularly large media client who was bleeding money on Wordpress.
3 engineers arrived - fashionably late. We explained them the situation and all we wanted from them was some GCP offering that would cure our woes and one that would cut our bills. The senior consultant - and presumably the only tech guy (rest seemed to be salesy) wasted our time like a used car salesman - he didn't even understand Google's own product portfolio and recommended us to use something like Spanner - which was totally not the solution to the problem, not to mention, expensive.
My boss and I left the meeting pissed off and he told me - "Neya, you probably know more about the product portfolio than these guys. Let's leave". That weekend, I went with my tried and trusted favorite Db - PostgreSQL - CloudSQL with a custom Elixir middleware based an old CMS I wrote a decade ago. After some trial and error, the solution worked flawlessly (and still does till date on auto-pilot). My client still has the lowest cost in the region - 1/3rd the cost of their competitors...7 years later. Back then, there was no vibe-coding, no AI, no auto-completion. Just pure thinking and experimentation.
All this just to say I agree that the new guy sometimes can make the best solutions to a problem and not always screw up. I always listen to new hires these days (now I'm a fractional / CTO) because you never know who could pull off that 1/3rd cost cutting framework move.
10 replies →
> Later my boss told me not to do that again because it caused havoc with schedules and such.
Did you talk to anyone about your plans before you brought in the demo or let them know they were solved problems? Often these sorts of reactions come down to your boss not wanting their team to lose their jobs because of the perception that it can all be handled by one person who's happy to work weekends.
28 replies →
Oh I have seen this story and was the one who caused this story when I was younger.
In a lot of cases the "new guy" thinks its an easy software and does it on his free time and thinks he did a great job.
In reality the specs are never 100% done correctly. The "new guy" misses some edge cases everybody but him knew because its just company knowledge. A lot of info in the specs was missing since they are not complete and so on.
This over the weekend never works in the long run. The ORM worked for all the happy path and written down cases but then you have cases were the ORM just is not good enough or fast. So you start to add strange code to work around the ORM. The same for the web framework or the validation lib.
To me the author of this comment sounds like the typical "Freelancer" coming in into a company knowing everything better then all the people and then leaving after a few months and now everybody else has to deal with his code.
4 replies →
> Sometimes the new guy can make a better version themselves over the weekend, not because they’re a supergenius, but because they’re not hampered by 47 teams all trying to get their stamp on the project.
In 99.9% of cases what seems to be "the better" version is better only for the "new guy" or rather his ego.
Those 47 teams hampering doesn't necessarily mean a bad thing, and more often than not actually well justified "stamps".
You only understand those things when you turn from the "supergenius" into an owner who have to take care not only of numbers on screen, but also security, interfacing, management and so on.
Or you don't turn into.
6 replies →
Take this idea and bring your own validation library and forms and UI components to the next job, and you've described what I do. And then you have real lock-in.
3 replies →
Yeah, that definitely happens.
But I don't think Claude Code is going to prevent an org that thinks they can prompt their way to a replacement for all their SaaS from having internal political bickering that makes them end up with a extra-shitty mega-compromise to try to make all the internal stakeholders happy.
If you've got no vision and no taste, you need to find a vendor who will protect you from screwing up your internal processes and tools.
Internal tools teams have rarely cared much about UX or the day-to-day experience of their poor users. The quick-and-dirty internal-prompt-based one is likely to similarly be unimaginative and unintuitive.
Isn't this agreeing with the parent? If Django were the B2B SaaS product, you didn't vibe-code Django, you just used Django. You aren't responsible for maintaining Django itself.
1 reply →
> feel free to doubt.
I don't doubt you at all.
I once worked on a project that was a new sign-up process for a large retailer (~90k employees at the time, four figures worth of outlets, quite a few billions in turnover).
There were about 60 people across, I can't even remember, maybe 10 teams that I knew about. One of the managers there assured me that the true participant figure was closer to 160. I was agog.
This project took months, and it had been going for months before I joined. And, as you can imagine, with that many people involved over that sort of timescale, there was turnover: people, sometimes key people, would leave and be replaced - sometimes by others who'd already worked on the project, but sometimes by new people - with the expected disruption that causes.
It was two web pages, plus some back end plumbing across multiple systems.
But, in the grand scheme of things, nothing that complex. I reckon it was a handful of thousands lines of code total across the full stack. I was mostly on the database side and, IIRC, I wrote a few hundred lines of SQL in a couple of stored procedures (wouldn't have been my preferred solution building from scracth but, you know, we weren't, and we had to integrate with a legacy systems that had to keep working as expected).
Overall it took 8 or 9 months from kick off to shipping and I was involved for perhaps 3 months of that towards the end. I was on the call with dozens of other people whilst me and one of the other guys hooked the final pieces together and tested it end to end for the first time... and it worked. There was actual cheering.
Large enterprises are genuinely ridiculous.
Rule #1 of power: never outshine the master
I had a similar experience where someone that had prior experience with django while we were using sqlalchemy started to design a django-like ORM on top of sqlalchemy. Of course it took him some time to get working, was a hell to understand and extend and most importantly lacked support for basic features.
Fortunately it was limited to a small isolated service but I can imagine the damage on the long term if you continue that route on what becomes a huge monolith after a few years.
It feels like I have worked with you (though obviously not). Had same experiences to the T.
Maybe it's just something that happens to many in web development world, not-invented-here and all...
I was complaining about SQLAlchemy's insane quirks to a group from my alma mater and one of the grad students said, "Well, the solution to your problem is clear: Write your own ORM." and I had to explain that this startup does not want to get into the ORM-writing business.
1 reply →
Also, some people want to work on what's already familiar to them. If building a framework from scratch is what appeals to them, they'll do that even if a framework already exists. Busywork to look productive.
I resonate strongly with this story. I’ve seen three people teams get in one month where SAP could not in year, and also let and witnessed incredible number of total fakers in SaaS enterprises.
Big corpo may be too big to fail, not so sure about their whole cohort of partners and fakers.
We are certainly closer now to being able to prototype and go to market faster with a product. In one weekend is a little much but I think its hard to deny that building will continue to expedite. What most developers don't think about is that the marketing, sales, customer service are all non-trivial parts of the business/product and all require legwork that is more than just sitting at an IDE. The nail in the coffin is that the data is a large part of company moats, and new products need time in the market to get that. Migration is also a long process and risky...so to get customers, a newcomer needs to provide way more value than what the incumbent gives.
I imagine you're going to have people trying to automate the whole GTM lifecycle, but eventually the developer that thinks they can bootstrap a one man enterprise without actually doing any kind of social interaction will run into a wall.
> We are certainly closer now to being able to prototype and go to market faster with a product
Absolutely. But this begs the question that businesses want to also sign up to maintain whatever product they've built, on top of their core business.
"Service" is the word that people seem to be forgetting in SaaS. If you roll you own, all you have is software.
1 reply →
In any usable product making a product is like 20% or less. Enter compliance, security, payments and a million other things.
Even if you can build it in a day B2B SaaS will continue to prosper because they sell peace of mind, reliability and compliance. Not features.
Also due to economy of scale it will always be cheaper to buy something from a vendor that sells it to many clients than to DIY it.
13 replies →
> We are certainly closer now to being able to prototype and go to market faster with a product.
What are the higher-order effects when anyone can do this, and *aaS becomes a market for Lemons?
27 replies →
I sure hope I never have to hunt down any GTM options myself soon and I can tell the AI to do what it should be doing. However AI adoption may be getting slowed down by profit motives because what Google should have already been doing is letting me git clone the entirety of GTM with all its configs to a local folder so I can treat it like code because it is code. The difficulty with AI adoption will be to make all products be like this so they can interact on a code level instead of me having to press buttons in different UIs to make thing shappen. E.g cloudflare should be letting me git clone its entire config, everything I did in the dashboard, to my folder too.
> bootstrap a one man enterprise without actually doing any kind of social interaction will run into a wall.
But you are not limited to only using LLM for coding.
I agree that marketing and sales is as important as product and technology, but they are not necessarily safe.
> We are certainly closer now to being able to prototype and go to market faster with a product.
Prototype maybe. Go to market maybe not so. It's giving false hope. You're just taking more shortcuts with prototyping.
This vibe-coding-will-replace-SaaS insanity is the new crypto-will-replace-fiat-money insanity.
It's shocking to me how prevalent this "who needs Salesforce when everyone can just vibe code their own CRM from scratch in a day" narrative has become in the business press. Like, what???
9 replies →
I don't think vibe-coding will replace anything. However, what if AI tools can make skilled developers more productive, particularly at simple tasks in unfamiliar environments? You could see that reducing the engineering costs of simple utility applications. There are tons of pitfalls that many here have pointed to but also maybe opportunities to do things that wouldn't have been cost effective.
Also: In my life the easier it has gotten to create and run software, the more software people have wanted and the more they have been willing to spend on it.
1 reply →
People really seem to believe that code is the only thing you need to make a SaaS company. It's like thinking a line cook is all you need to open a restaurant. There are so, so many other components to running a business.
4 replies →
I don't think it will replace SaaS but I do think it can replace the need for a lot of the consultant work that goes around configuring and integrating the SaaS. It will be much easier to have a spec that defines how things need to be configured and the machines can implement it (using the SaaS as tools). Frankly this is the most annoying part. It's not that the B2B stuff can't do whatever, it's that it never gets implemented in ways that aren't a pain in the ass because it's all handled by people who aren't actually using them.
I really don't think it's not going to become "these prompts are specs" and then you have processes of reviewing implementations. It's one thing when you have randos building stuff and they leave etc. Having stored prompts and managed code that uses tools is a different beast.
1 reply →
No serious programmer "vibe" codes. I admit creating SaaS may not be feasible with current infrastructure but you can't ignore the insane jump in productivity that these tools can offer with the right scaffolding.
1 reply →
There is still a narrative here. Lots of SaaS recurring revenue is built on value-add features that can be easily replaced.
What I struggle with is developers wanting to leave platforms like Datadog for open source equivalents that need to be self-hosted.
I hear all of the cost savings benefit, but I never see the team factoring in their own time (and others time) needed to set up and maintain these systems reliably long term.
Something IC’s at company often struggle to understand is the reason why companies often prefer to buy managed solutions even when “free” alternatives exist (read: the free alternatives are also expensive, just a different type of cost)
My log bill for Google cloud log would be like 30k. For splunk I like 80k. I self host for 1.5k per month. Spend maybe an hour a month? Easiest money I ever made.
18 replies →
Do they actually not understand that? They might just be fine with a system that makes them more useful.
How do you calculate the time spent on an internal tool like this, actually? (I’ve never been in management). Realistically your team inevitably will have some downtime, maybe some internal tool maintenance can be fit in there? I mean it obviously isn’t fully “free” but is also shouldn’t be “billed” at their full salary, right?
5 replies →
Because most of them arent trained to think economically... how many people on the planet do you think are aware of the notion of opportunity cost?
If there's ever any use case to leave an expensive SaaS for self hosted, you can find it at datadog
>the free alternatives are also expensive, just a different type of cost
Not if you hire reasonably competent people. These days for vast majority of FOSS services all you need is an ability to spin up a VPS and run a number of simple Docker/Podman Compose commands, it can't be that hard.
3 replies →
I'm sorry but the amount of companies that need something like DataDog is quite small compared to their 30,000+ customer count. Maybe 5,000 companies on Earth truly need something like DataDog, 80% of their customers would be perfectly fine with a self hosted instance of grafana.
Using an open source self hosted solution should be the industry standard, encouraged position, by default. Our industry does not gain overall from using DataDog but only from truly open source solutions that utilized AGPL licenses that allows everyone to move forward together + share lessons together + contribute together toward a common goal of better observability.
Why are we acting like it's hard to set up? This isn't the 1990s, it's 2026. Tooling has gotten quite good over the last decade.
Also corporations stupidly spend money all the time, they over spend too. I recently left a company that was paying SalesForce $10mil a year in licenses when only 8 people in the entire 3,000 person company was using it. I doubt that was the only single instance across our industry too. There is a massive amount of waste and graft in enterprise sales.
I honestly doubt it if you replaced grafana for 10,000 DataDog customers they would notice the difference.
6 replies →
& the counter-argument is those SAAS apps being killed by A.I are growing revenue 20%+ YOY
people who write this BS - one never don't understand SAAS fundamentals, they only see what's on the screen and forget the complexity that lives on the backend - forget the costs of running such a SAAS
before it was low-code will kill SAAS, then Visual UI builders, now its A.I
just like it was before that crypto will kill Trad-Fi
people who say these things - have tied their identity into it so they whole-heartedly believe the bullshit they say even though reality doesn't match
to anyone curious read the 10k (Annual Report) of any public SAAS - Salesforce | Workday etc - people should admire these companies for the machines / ecosystem they built - and also learn the good & mistakes to avoid i.e the bad
those annual reports tell you how the revenue generation machine works, how much revenue is expected 2+ / 3+ years from now - their weaknesses | headwinds and also tailwinds - how those companies grow and continue to grow etc
Here are the most recent 10-Ks for Salesforce and Workday.
https://www.sec.gov/ix?doc=/Archives/edgar/data/0001108524/0...
https://www.sec.gov/ix?doc=/Archives/edgar/data/0001327811/0...
SaaS companies are effectively both "software maintainers" and "price gouging middlemen" at the same time. The difference between the bid and the ask for SaaS is part of a simple math problem for whether the company should try to create their own version of the software they need. It may be the right decision, it may be the wrong decision, but it will be the right decision for a non-trivial number of firms. And that means SaaS businesses will both lose customers and have downward pressure on their margins. That means valuations of B2B SaaS firms go down.
My career occupies a weird middle ground where, for 20 years or so, I've catered to smaller businesses that need bespoke solutions (because the SaaS available doesn't conform well to their business logic), but don't have the scale or desire to build and maintain software in-house. Sometimes these are slapped together in a weekend, if that's all that's needed. But in most cases they still become ongoing improvement and maintenance projects for me.
This niche position has had some interesting ramifications for them and for me. They clearly incur a lot of technical debt once their business relies on bespoke software. On the other hand, they own the software and can get an immediate response or new feature or upgrade from me, limited only by my time. And in the end, this ends up saving them time and money. It gives me a permanent and unending flow of work. But if I die, they're pretty screwed.
One reason I don't vibe code things even now, even simple components that could easily be vibe coded, is that I remember and know where everything is, every function or line of code that might be causing issues, because I wrote it myself. I know right away where to look for a query that might be throwing errors after a database upgrade, for instance.
As a manager I assume you would probably not want to go down the road of hiring someone like that, but for companies of a certain size it's an acceptable compromise. However, I wouldn't want to hire someone like that myself unless they were extremely reliable and didn't rely on AI to write any of their code.
People sometimes fail to appreciate the value of KNOWING the system inside and out when it comes to diagnosis and troubleshooting.
Observability is great, dont get me wrong, but past 3 to 6 months of work on the same thing...I can almost beet the observability tools in timetoresolve.
This sounds great if you get on well with your clients. You must be an effective networker and at sales. How do you bill, and how do you price your services?
1 reply →
So you think this downturn will be short lived?
When management realise that the vibe coded projects are not maintainable, SAAS will be as popular as ever
It seems that current advantages would compound with AI. I.e., if I am making a SaaS for Popsicle stick makers today, why I am disadvantaged with AI vs a new competitor in the space? I guess the hypothesis is the Popsicle stick maker will vibe code all of the software that they need instead. For that, we need significantly better AI than we have today - perhaps something like a 1000X improvement. Basically, this is a world in which non-technical grandparents can vibe code anything that they want. This means, it understands what you want without you being able to articulate it well in the first place.
8 replies →
I don't do tea leaves so I wouldn't commit to that, particularly because I think SAAS was oversold in general even before LLMs came out. But I think the idea that the industry as a whole will shrivel away just isn't feasible, even if there is a correction.
3 replies →
I dont even blame management. I believe most of them are well-aware that much of what is going on right now is pure hype.
However, they dont have a choice. The sentiment of shareholders is that they want their cash (yes it is their cash that managers re-invest on their behalf) to be invested in AI-related projects.
So...... you get what you get, and investors will get what they deserve. But they will still blame the management in the end ;)
In many cases, it's not a downturn, just a return to reasonable valuations. Other sectors should follow
All of the hype surrounding AI will subside when a SaaS company eventually deploys a moltbot version of their software and the company is driven out of existence due to the chaos that ensues.
2 replies →
I've seen some bad some SaaS that management insists is an integration they need and no one else provides. I can easily see some vibe coded projects replace them.
They will magically realize this when their huge bonuses will be tied to something longer lasting than last quarter/year performance on some very narrow metric (which has nothing to do with sane stuff like adding long term value to some part of the company).
They are not stupid, far from it, most are (very) high functioning sociopaths. And out and up there its everybody for themselves first.
A huuuuuge part of why companies, especially public companies, especially those in regulated industries like healthcare and finance are willing to pay eye-watering sums of money for a SaaS app that you could get an MVP up in a few weeks time from a competent team with no AI is that they need a phone number to call when something shits the bed at 2 AM on a Wednesday. They need support SLAs without the payroll and headache associated with it. They need someone to sue if things truly go tits up.
Moving SaaS apps in house is a great way for a VP to get a fat bonus or a director to get a promotion but I have to imagine it keeps the CIO/CTO up at night unless they're fully asleep at the switch.
Yep! We sometimes have a choice between the gold-standard and commonly updated open source solution to X and a two-bit hacked together proprietary solution that has 24/7 support at high cost...and we choose the one with support, because that's what our audits basically require. Because then we can say "yes, it's still within the support contract, we have an escalation point".
In my career I've deployed close to 100 different saas products at enterprise level and can tell you that most of the current crop are slapped-together half-finished dross with a huge sales and marketing team.
In the time it takes to deploy semi-bespoke saas, or while waiting for the current licence term to expire it would be very easy to develop a more suitable and much cheaper product in-house, this was true before AI tools and doubly so now.
The problem is right now management is not only insisting on their team vibe-coding bespoke replacements, they’re avoiding paying for other SaaS because they can vibe-code their own replacements, often themselves, and they’ve lost sight of that they probably don’t want to be responsible for it.
Besides rampant failures in communication and skills allocation, wild U turns of requirements were (sometimes, not even real business requirements) were holding back corporate environments doing a decent job.
With AI, I can only see the rate of such changes sky rocketing due to expectations wildly misaligned with reality. Hence we are unlikely to see any meaningful improvements.
This is an incredibly broad statement that just isn’t true in a million cases. Last co we migrated all of our observability from Datadog to Grafana/AMP because it’s much much cheaper. The vendor can charge some premium over the cost to build/maintain but not infinity. SaaS is going to have to get dramatically cheaper to compete with the lower cost of building your own.
what if this time it's senior developers and they actually can slap something together better then the expensive SAAS offerings?
what if the expensive SAAS offering is just as vibe coded and poor quality as what a junior offers?
You're not considering opportunity costs and buyers vs. users.
If your senior developers can slap together something better than an expensive SAAS offering you want them directing that energy at your core products/services rather than supporting tools.
And the people deciding to buy the expensive SAAS tools are often not the people using them, and typically don't care too much about how crappy the tool may or may not be for doing the job it's advertising as doing.
2 replies →
Clubbing all saas products together just means you can’t really have a productive discussion. Saas products are on a spectrum of quality, from amazing (stripe, datadog) to terrible (fivetran, github). Its upto you as a user to make a call as to which will serve you best, what you should focus your limited resources on etc.
> what if this time it's senior developers and they actually can slap something together better then the expensive SAAS offerings
A typical SaaS customer will use many pieces of software (we mostly call them SaaS now) across its various functions: HR, accounting, CRM, etc. Each one of those will have access to the same pool of senior devs and AI tools, but they will pour more resources into each area and theoretically deliver better software.
The bigger issue here is the economics of the C-suite have not changed here. Assume a 100 CPG company uses 10-20 SaaS apps. Salesforce might be $100k/year or whatever. 1Password is $10k. Asana $10k. etc. They add up, but on the other hand it is not productive to task a $150k employee with rebuilding a $10k tool. And even with AI, it would take a lot of effort to make something that will satisfy a team accustomed to any modern SaaS tool like Salesforce or Atlassian. (Engineers will not even move off Github, and it's literally built on free software.)
That's before I get to sensitive areas. Do you want to use a vibe-coded accounting system? Inventory system? Payroll? You can lose money, employees, and customer perception very rapidly due to some bugs. Who wants to be responsible for all their employee passwords are compromised because they wanted to save $800/mo?
Then, the gains from cutting SaaS are capped. You can only cut your SaaS spend to zero. On the other hand, if you have those engineers you can point them at niche problems in your business niche (which you know better than anyone) and create conditions for your business to grow faster. The returns from this are uncapped.
TL;DR; it's generally not a great idea to build in-house unless your requirements are essentially bespoke.
3 replies →
To the first question, if your senior devs can do that there's almost certainly something more directly valuable to your business they could be doing than solving a problem your vendor has already solved
The second question is a valid one, and I think it will somewhat raise the bar of what successful SAAS vendors will have to offer in coming years
1 reply →
Nice what ifs, but not valid so far. I get the motivation to think/hope so, but thats not the proper business world right now where big money are. Maybe next year it could start becoming true but then market will be a bit different too
You're ignoring the biggest part of SaaS as far as management is concerned.
There's a large, stable entity that management can sue if something goes very wrong.
There are of course exceptions to every rule, and I'm sure some companies have been successful in building their own in-house tooling.
At the end of the day these decisions are all series of trade-offs, and the trick is understanding your requirements and capabilities well enough to make the right trade-offs.
It that works, nobody would be using Jira anymore, because people would just use a competitor that's cheaper or vibe code their internal Jira tool.
Somehow that has not happened yet in 2026.
6 replies →
I think there’s also the classic “I can build zoom in a day” - they get video working between two machines. But it’s the last 80% of the app that takes 99% of the time. Nerds don’t see the whole product, just the happy path of a wee technical challenge.
This may be true pre-LLMs, but I think you need to account for the baseline build-vs-buy tradeoff shifting.
Companies in most cases don’t want to build SaaS because it is expensive to hire engineers to do it, not because they are allergic to owning teams.
If in-housing becomes substantially cheaper than the alternatives then companies will adapt.
But even if the new equilibrium is to hire a contract dev shop to build something custom to keep avoiding responsibility, this would have the same impact on SaaS.
So I’m pretty skeptical of this first-principles prediction expressing right level of uncertainty.
You’re forgetting the companies that already had developers.
Whose job had been maintaining a single internal system but had never had the bandwidth to expand their focus.
Companies like that are the ones spending millions a year for large one size fits all SaaS products.
The same cost savings could be captured by SaaS. Potentially not by incumbents, but by up and comers.
If you can spend $10K/year to keep your in house one alive but $5K/year on the new SaaS option, you stop building your own again.
3 replies →
I totally agree about the management reluctance to just own everything in house.
But I think it’s plausible that SaaS companies will be easier to start with AI coding, and with lower costs (thanks to AI) they will be able to get into the black with a smaller addressable market. So each one can have a different mix of fewer features, for different segments of customers, at lower prices.
The result would be a loss of pricing power by the incumbent do-everything big guys: no more baked-in 10% annual increases. Which is still a pretty big change in their economics. And therefore valuations.
The companies that already have a strong in-house team will greatly benefit from AI. Many of those who don't are in that situation because managers have PTSD from so many failed projects. Half of all projects fail. That's a lot of emotional trauma.
This was all possible pre-AI. The reasons that some Saas companies win have nothing to do with how quickly or cheaply code can be written for the Saas.
I can't count the times I've told clients and prospects to _not_ hire us to build something they wanted. Because they could just use off the shelf solutions that were cheaper financially, at least in the short to mid term, and much, much cheaper in terms of opportunity costs. I struggle to put even billed hours into something that doesn't make sense to me.
Of course some overdo it. I've seen companies with more random SaaS tools than staff, connected with shaky Zapier workflows, manual processes, or not at all. No backups, no sense of risks, just YOLOing. That's OK in some cases, in others really not.
I suppose it does need some engineering thinking to find the right things and employ them in a good way. Unfortunately even developers commonly lack that.
I'm a manager too, but I'm also the new guy pushing the solution to a human problem: work management. SMAR, ASAN, MNDY, etc. Not only do people not want to be responsible for it (and in some cases simply be "not responsible"), not only is the internal solution "too time consuming", the only answer thus lands on hiring external consultants to implement and maintain massively-overkill-$olution$ in $aa$ like CRM, NOW, etc. which as you know, do not solve the same problems as the aforementioned SaaS.
"Now that I'm in management, I 100% get it." 100% and win or lose I am still going to fight it...
And since you became management, you have refered to software only by the stock symbols of the owning companies?
But this time management has to justify its AI spend.
We've been through cycles of outsourcing and in-housing skill before. This seems similar but for tools and services. Maybe we'll see internal teams taking the reigns back to replace bad-fit SaaS products.
There's still a lot of risk associated with in-housing though (perhaps more than before). That means the real opportunity is for new, leaner B2B SaaS businesses to step in, especially if there's a displacement effect from seeing internally built prototypes of expensive subscription software.
There's that and then there are companies spending 100k on a software suite just to use that 2 features. So now one of their junior dev solves it and becomes a hero. The truth it always somewhere in the middle.
I think the pressure on SaaS margins won't be from customers vibing their way to Figma or DataDog but because gen AI will bring a lot more credible competition in many segments. DD and Figma are probably awful examples because those companies are constantly pushing the envelope, but there are a lot of rent seeking SaaS providers that are going to be in for a rude awaking.
The difference between a vibe-coded prototype product, even a good one, and an enterprise SaaS platform is the difference between a Lightning bug and a Lightning Bolt.
To be fair, re-creating the SaaS solution that simply replicates the features they see can often be done fairly simply. However, there are generally a whole lot of things under the surface. Then there is the whole hosting and maintaining the system, which is its own problem.
Also, management doesn’t have time to fully understand it, which means they need at least one employee who does. And that employee now has leverage.
> "just do the parts of it we actually use".
25 years here. You can absolutely do this. Most software is orders of magnitude more complex than it needs to be.
The junior programmer you are talking about who wanted to rewrite it in a weekend tends to come back with a working program, not empty handed.
I've seen this happen with both juniors and seniors. They do come back with a working solution /for the happy path/. Because the happy path is easy. It turns out that most of the complexity sits in the unhappy paths.
5 replies →
I agree. Just because you can buy some piece of software doesn't mean you should -- there is a lot of software that exists just to sell more consulting hours and will never fit the business. It's actually not hard at all to code and maintain much simpler alternatives.
Actually having to support multiple businesses with commercial software is hard. I've written a ton of custom software that far surpasses the capabilities of commercial offerings but if were to turn that into it's own commercial offering it would be large undertaking.
Yes, I didn't really doubt the developer could do it, the problems are:
1. That's not a great use of the developer's time, and
2. anything in-house increases our training and support costs
9 replies →
If the management is the one actually paying for the software from their own pocket (founder), the tables turn. There are millions of SME owners who are forced to pay for B2B software just out of necessity and not having resources to build it in-house.
AI could change that for good.
I work at a smallish public tech company and while this may be true at some companies, it's not true at all of them. We have almost no SaaS vendors. If we do have to buy software, we're almost only interested in On-Prem.
It almost always devolves into some all encompassing ERP that is meant to solve the needs of all parts of the business and save millions in licensing costs, and we all know how well that plan goes.
Software without support is useless. In the business world, what's being bought isn't code—it's a solution to a business problem, with a throat to choke if things go wrong.
I guess one difference is now you can ask an LLM, trained on an open core product, to generate a license-free version of it.
management simply doesn't want to be responsible for it
That sounds dysfunctional. The purpose of management is to manage risk, not to avoid it. A proper manager would be able to quantify both the risks and the costs, present those figures in an easy overview, and then be able to defend their decision (or advise higher management) using that.
My firm has partially transitioned through this curve. We went went from "fully externally supplied" systems, to an architecture that combines "externally supplied" (core functionality) with "low code" about 6 years ago. I would argue (as a financial manager) that that lead to a more flexible and more affordable architecture. A funny mixed bag problem arose though: the curve of business demands grew harder than that of IT-delivery. So IT delivered more value, but business keeps demanding a faster pace. If I project this line to the future AI will most certainly harm our external suppliers. We keep getting better at DIY development and "low code" will transition towards "no code". Not really "no code" of course, but DIY IT developed tooling.
The age of the business developer has re-arrived.
For the first time in my career, I can point to multi million euro external suppliers, tell my environment "that's basicly an API + authentication from X to Z, let's develop that ourselves" and get a response of "When" instead of "No". B2B SaaS is toast in my perspective, as are boutique firms delivering solutions + consulting. I can create a million euro team easily (that's like five developer years), if they deliver a successful insourcing. And now I feel like writing MBA-slop, but's it's all about growing your IT maturity. All insourced code is future maintenance expenditure. You need to balance that to the benefits.
> All insourced code is future maintenance expenditure. You need to balance that to the benefits.
I love this perspective. I feel like the pendulum has swung too far back to "it's easy to build, it'll be easy to support". But to be fair, it was probably too far the other way a few years ago: "it was easy to buy, it'll be easy to have them support it".
Other than trial and error, how do you think about pricing out maintenance costs for insourced code vs purchased functionality?
Building a project is very different than building a product as a billing engine.
> management simply doesn't want to be responsible for it
The problem with this kind of thinking is that it strips away all nuance. At some point you have to be responsible for something ... otherwise you don't have a business. You are simply a wrapper around your SaaS providers and tightly coupled to their success. The key is knowing when to offload and when to keep it in house. Quite frankly, your average weekend MBA VP simply doesn't have the expertise to make these kinds of decisions. This is why so many VPs exit before things get bad.
> At some point you have to be responsible for something ... otherwise you don't have a business.
Uh, yeah? No kidding. That's why you focus on your core business. If your core business isn't "writing a new and better Jira", don't write a new Jira.
Reminded me of big balls and DOGE. Didn’t we vibe code our entire government by now?
Your profit margin is my opportunity.
Think about it differently - let's say a free OSS product can be installed and you can use ALL features except for LDAP (because that's the paywalled portion that requires you to buy it for $25k / month.)
Well, with claude, you can download the code, tell it to implement LDAP authentication, and smile all the way to the bank. And for said fortune 500 company, employing an employee to spend 100% of their time maintaining the app at 10k per month is a 15k savings! And because it _doesn't really take 100% of their time_ it's really only like $500 per month? And to be completely honest, how man times did you get Jira to fast-track your issue?
I get it however, the manager angle. It's still a distraction. But the article being referenced still shows revenue going down.
There's definitely a lot of cope in here, mostly because SaaS is keeping them employed... be ready, the crush is "almosthere".
> an employee to spend 100% of their time maintaining the app at 10k per month
If the cost to the company is $10K a month, the developer's topline salary is $60K, which is going to be a hard hire to make.
And, again, if they can integrate LDAP with an existing software package at that price point, I want them doing something more valuable than that.
sorry, what do you mean?
1. Enthusiastic employee (vibe-)codes a replacement for a turnkey SaaS product that the company uses.
2. Company uses it, maybe even starts to rely on it for important business operations, and for a time the employee supports that app.
3. Bugs creep in, feature request pile up.
4. Employee either leaves the company or moves on to another project.
5. Pain
13 replies →
If I understand correctly many organizations will not develop original stuff internally, because nobody internally wants to be the one is shouted at if something goes wrong.
2 replies →
If software already exists that does X, X is a solved problem. You didn't hire a developer to solve already-solved problems.
2 replies →
[dead]
I like to bring up JIRA example. You could replace it in-house yeah it is just tickets with statuses. /s
But then keep in mind one who built the replacement will become the owner of an application that business doesn’t want to pay for and that person will be cost center for the company.
That person better get marketing and negotiating skills that Atlassian has on board because that person will be responsible for the app and will not be getting salary increases for working on something that is not core business of the company.
Even if you can make LLM to do the app for you.
You guys keep using services like Jira, Salesforce, Stripe, Datadog, etc. While those are definitely the biggest names, I don't think people are referring to those SaaS platforms as the ones they will replace or try to build an inhouse version of. It will be things like ETL pipeline services, data scraping services, maybe some internal analytics SaaS. The niche things that cost a lot because they’re in a sweet spot where only a few people need them, but no one used to have the resources to build them in-house. So, when the salesperson called and offered a perfect solution to their problem, they bought the service. Those are the ones that will be more targeted for in-house solutions.
4 replies →
I'd actually say the opposite is the case. B2B (even SaaS) is probably the most robust when it comes to AI resistance. The described "in house vibe coded SaaS replacement" does not mirror my experience in B2B at all. The B2B software mindset I've encountered the most is "We'll pay you so we don't have to wrestle with this and can focus on what we do. We'll pay you even more if we worry even less." which is basically the opposite of...let's have someone inhouse vibe code and push to production. B2B is usually fairly conservative.
Reminds me of a blog post a while back saying that gigabit fiber at home would lead to everyone running their own email server.
There was no chance that everyone would be running their own email server, but if it wasn't for the lack of IPv6 adaptation a plug and go home email server solution would probably see a decent amount of use. I'd bet we'd already be seeing it as a feature in most mid-ranged home routers by now.
17 replies →
The email server bit directly correlates, too. Will everyone vibe-ops their own email server using AI? Of course not.
What a weird take. I was running my own email server 25 years ago on a 512 kbit ADSL line. No problem at all, would even be enough bandwidth today for most messages.
(Back then email still worked from residential IP addresses, and wasn't blocked by default)
I agree with you. In B2B SaaS you don't sell the software, you sell your expertise in a specific domain and the responsability you take for owning that expertise. The fact that the development costs are nearly zero will make them more valuable and more protifable
Development has always been a small fraction of B2B SaaS expenses after the first couple of years anyways
B2B is a large corp is like you describe, but it's very different in SMBs, and there are many, many more SMBs.
My experience is that SMBs are generally not run by people who feel confident doing any kind of self managed IT.
No amount of LLM usage is going to change them into full stack vibe coders who moonlight as sysadmins. I just don't see it happening.
Not until, that is, a new generation, that has grown accustomed to the tech, takes over.
Until then the current SMBs will for the most part fulfill their IT needs from SaaS businesses (of which I think there will be more due to LLMs lowering the barrier for those of us who feel confident in our coding and sysadmin skills already).
4 replies →
I'm considering SaaS replacements with in house code in situations where my general thoughts are "how can this possibly be the pricing for this?" which is not uncommon.
Well before vibe coding, tons of open source software existed (and exists) to replace SaaS. With lots of features and knobs and real communities. But I still often pay for SaaS because managing it is a headache. Some human has to do it. I can pay the human or I can pay the company. I really don’t see how vibe coded toys can replace real battle tested SaaS products. A better explanation is the bubble in PE ratio is deflating and it’s happening all over, regressing to the mean. AI is a convenient explanation for everything
8 replies →
Maybe you are right and the companies do want to pay and not worry about these problems. But now they have a lot more SaaS options to chose from. The incumbent companies like Salesforce and Atlassian have less of a moat. Maybe they'll keep the power users but if a customer is only using 80% of the feature set there is new competition. Competition might come in the form of a startup but it can also come from existing SaaS companies expanding into adjacent domains. Canva now does docs. Notion does email. etc
Also, it is my experience that exec and boards favour safe and well known B2B partners over in house. It's a more publicly defensible approach that gives them an out if things go wrong and shareholders get upset.
For big corporations at least prices of SaaS are rarely an issue. Issues are: we don’t have the time to introduce a new tool, what about our processes, we don’t have the right people.
So how much Constellation Software stock are you buying since the market seems to think they are dead in the water after a 50% drawdown?
hard disagree, several b2b categories are going extinct because AI just completely replaced them.
I mean if we want recent examples just look at tailwindui since it's technically a SaaS.
> we want recent examples just look at tailwindui since it's technically a SaaS.
This is a terrible example. Show me someone ripping out their SAP ERP or SalesForce CRM system where they're paying $100k+ for a vibe coded alternative and I'll believe this overall sentiment.
9 replies →
TailwindUI isn't really what I'd consider SaaS -- it was a buy once and download software product.
That means to keep making money they need keep selling new people. According to them, their only marketing channel was the Tailwind docs, AI made it so not nearly as many people needed to visit the tailwind docs.
If they had gone with the subscription SaaS model, they'd probably be a little better off, as they would have still had revenue coming in from their existing users.
> I mean if we want recent examples just look at tailwindui since it's technically a SaaS.
How is it in any way B2B? At most B2C + freelancers / individuals / really small SME.
It didn't have any clues a med/large B2B would look for e.g. SSO, SOC2 and other security measures. It doesn't target reusability that I as a B would want. The provided blocks never work together. There aren't reusable components.
Tailwind UI or now Tailwind Plus is more like vibe coding pre-AI.
Sorry but tailwindui is not a SAAS. There is no service or hosting. You buy a coded template once and then receive updates. It is totally not the same as a critical B2B SAAS that is running 24-7 on the vendor's servers providing real support and service.
TailwindUI unfortunately sits in a position of being an easy to disrupt business with current AI.
Now attempt the same with Zoom, I suspect vibe coding will fall down on a project that complex to fit the mental model of a single engineer maintained a widely used tool
Perhaps the case for premium CSS SaaS businesses, I guess (which seems particularly primed for disruption even pre-AI), but there are many more robust B2B categories out there that aren't literal code + docs as a service.
There is a paradigm shift but personally I like to zoom out a little:
It used to be that your new b2b product has to try and displace a spreadsheet. Now it has to displace an agent.
1 reply →
how dont people understand? if you have a VC funded b2b saas, you need to charge huge margins for the investors to get a return. now, small teams can vibe code a replacement and charge 90% less money. AI is going to kill saas margins.
i literally cannot understand why people keep repeating that non tech companies will build their own software, thats not the bear case for saas
Atlassian: surviving since 2002 because no-one could previously build a kanban board or project management app
7 replies →
Yeah.... The code isn't the hard part. That's not where the value is.
This hard part when you're doing in house stuff is getting a good spec, ongoing support, and long term maintenance.
I've gone trough development of a module with a stakeholder, got a whole spec, confirmed it, coded it, launched it, and was then told it didn't work at all like what they needed. It was literally what they told me... I've said 'yes we can make that report, what specific fields do you need' and gotten blank stares.
Even if you're lucky and the original stakeholder and the code are on the same page, as soon as you get a coworkers 'wouldnt it be nice if...' you're going to have a bad day if it's hand coded, vibecoded, or outsourced...
This has always been the problem, it's why no-code never _really_ worked, even if the tech was perfectly functional.
1 reply →
The reality is anyone generate useful code with an AI agent now. Dores in accounting can now automate all her spreadsheets in a single afternoon.
Not trying to hype AI, but we are in an interesting transitional period.
The accounting saas dores presumably uses doesn't "automate spreadsheets" as its core value prop.
related: i'm thinking these vibe coded solutions are revealing to everyone how important and under appreciated good UX is when it comes to implicit education of any given thing. Like given this complex process, the UX is holding your hand while educating you through a workflow. this stuff is part of software engineering yet it isn't "code".
[dead]
I, on the other hand, can't wait to fire every single B2B subscription we've got.
B2B SaaS is a VULN. They get bought out, raise prices, fail. And then you have extremely large amounts of unplanned spend and engineering to get around them.
I remember when we replaced the feature flags and metrics dashboards with SignalFX and LaunchDarkly. Both of those went sour. SignalFx got bought out and quadrupled their insane prices. LaunchDarkly promised the moon, but their product worked worse than our in-house system and we spent nearly a year with a couple of dedicated headcount engineering workarounds.
Atlassian, you name it - it's all got to go.
I just wish I could include AWS in this list. Compute and infra needs to be as generic as water.
If you're working at SaaS, find an exit. AI is coming for you. Now's a great time to work on the AI replacement of your product.
> And then you have extremely large amounts of unplanned spend and engineering to get around them
You get the same shocks with internal teams, just from other causes. And you have to manage them.
I'm sure you've only ever seen brilliant software created by internal software teams?
> And then you have extremely large amounts of unplanned spend and engineering to get around them.
I have no idea how you are spending "large amounts" of unplanned spend on Saas products. Every company I worked for had Saas subscription costs being under 1% of capex. Unless you add AWS, which is actually "large amounts" but good luck vibe coding that.
2 replies →
If you’re working in engineering, find an exit. AI is coming for you.
1. This isn't rooted in data but anecdotes "One Series E CEO told me that they’re re-evaluating the quarterly renewal of their engineering productivity software because they along with an engineer reimplemented something using Github and Notion APIs. They were paying $30,000 to a popular tool3 and they were not going to renew anymore."
2. These anecdotes are about tech startups spend, not your <insert average manufacturing business>. Nor or they grounded in data that says "we interviewed 150 SMB companies and 40% of them have cancelled their SaaS subscriptions and replaced it with vibe coded tools"
3. "Analysts are writing notes titled “No Reasons to Own” software stocks." - there is just one analyst saying this: https://finance.yahoo.com/news/no-reasons-own-software-stock...
4. Most of these SaaS tech stocks have been trading at all time highs...this smells of "explain something very complex with a simple anecdote"
EDIT: Oh lol, the author has a vibe coding SaaS offering...there ya go.
As someone who works in the B2B SaaS space...
- If our customers vibe coded better integration points for us, it probably improves our overall value to our customers.
- The software industry, especially startups, is such an insignificant portion of the market, its not really worth worrying about. But, I can tell you from experience, that even large software companies don't want their own developers spending much time on accounting, ERP, or HRIS systems and they "outsource" this to SaaS companies.
It would be amusing if somebody parodied this article titled "AI is killing speculative articles about AI".
Not what you are asking for, but the Onion has some AI tips: https://theonion.com/tips-for-using-ai/
Fav tips:
- Give the AI restraints, like “Don’t tell me to kill myself as part of this stir-fry recipe.”
- Fact-check any information provided by asking the follow-up question “Are you sure?”
- Offset your water footprint by not bathing for 72 hours after each use.
Exactly. This sounds like a barber advising which haircut to get.
That too on a 30k contract. That CEO likely sells some AI software and is under tremendous VC pressure to show AI investment.
Yes, and the content itself is clearly AI generated marketing nonsense.
Yeah, this is just long-form linkedin slop. He's thought-leadering to get you to get his (no-doubt slop-written) guides and do leadgen for his forthcoming saas.
sigh.
Yeah, also if a SaaS costs, 10k a year, I promise its not not more cost effecient to pull your 10k a month engineer off their usual work to build and then maintain some vibe coded slope everytime an edge case occurs.
Also many customers of SaaS have little to zero engineering staff, they are in construction, resturaunts, law offices ect. These takes are so assanine.
Is there a place on the internet where folks like yourself, who seemingly have a way to think economically congregate? I personally dont know of one, for which if I did, I wouldnt visit here anymore.
So many takes on here are so lazy and simpleton that when you go a few levels deeper all the flaws get exposed.
4 replies →
Even in companies that have SWE, do you really want to divert in-house SWE time to something as exciting as ... accounting rules and making sure your inventory is auditable? Or any number of the weird compliance things associated with most B2B software for a medium-size business?
I have few sticks in the sand in my thinking framework:
* writing code has always been the easiest part of building software, deciding what to do and what not to do is something else that takes forever sometimes
* there are several open source projects that can replace commercial SaaS and still people prefer to purchase commercial SaaS. These are available immediately, deployed immediately etc etc.
* along the same line, some of those open source projects offer self-hosting and cloud version: I would always personally go for the cloud version because in a small team I don't want to operate something that other people built and I don't know how to operate. That's not my job not my team job
* people are underestimating how draining is operating and maintaining software, which is something that goes beyond the adrenaline rush you get after "building" something with Lovable or similar tools. Also, I find it extremely easy to get 80% done quickly but excruciatingly slow to get things done right.
* I still see huge value in using tools like Lovable to build a working prototype and validate assumptions so that you get quickly build the right thing right solving the right problem in the right away avoiding waste
* camcorders have been around for ages but you don't have millions of directors around just because you make a tool more accessible
* same can be said for other things like restaurants, where sometimes it's more convenient (although expensive) to buy vs build.
>people are underestimating how draining is operating and maintaining software
Yep. Many SaaS have an edge because they factorize the struggle of many customers, if a SaaS has 1000 customers, each customer vibing their way into a home-built solution will require dedicated efforts at maintaining it. Even with AI, those efforts aren't negligible.
Many companies don't even operate any IT infrastructure, cloud or otherwise, themselves, beyond office connectivity, AI replacing SaaS will require someone in charge of that at the very least.
> camcorders have been around for ages but you don't have millions of directors around just because you make a tool more accessible
tiktok alone has 1.5 million directors! it's just that we call them creators now.
the meaning of the word director has changed, that's all. but professional roles shift in meaning all the time. a computer used to be a literal human doing arithmetic. an engineer was someone who designed war machinery. being a doctor used to mean teaching at an university.
human beings are natural tool makers. we always have been. the frontier material to manipulate where the most advanced engineering happens constantly changes: stone -> bronze -> iron -> ink (descartes) -> steel -> silicon -> javascript (YOU ARE HERE).
notice how each step is an improvement/abstraction on top of the steps that came before it. some say english is next in that chain. i honestly have no idea. all i know is there will always be The Next Thing and it'll be much nicer to work with.
I agree with your larger point, however:
> camcorders have been around for ages but you don't have millions of directors around just because you make a tool more accessible
That’s exactly what happened. There are more filmmakers than ever now due to the accessibility of cheap cameras, then digital tools, including affordable HD cameras. Especially once the DSLR revolution took off circa 2011, which enabled budget-constrained aspiring filmmakers to use prime lens sets rather than fixed/built in zooms on cheap camcorders. With proper lighting they could actually make something look pretty damn cinematic. The entire industry has radically shifted in the last 15 years in particular due to these changes, but it started to shift around 2000 if I have to assign a particular year to it.
When you had to shoot on film stock, which was expensive and had a whole processing pipeline that one couldn’t reasonably do at home, there was a much thicker barrier to entry. You basically had to go to film school or get into the industry before you could start making your own stuff. Hell the Duplass brothers started out on crappy camcorders. Now? A smartphone, some cheap LED’s, a basic computer, you can really make something.
My point was along this line: writing code is different than building a product as much as recording a video is different than telling a story with a movie (long and short that you want).
Do everyone has the capability to build a comprehensive set of features that we call a product to solve problems that people or business have in their life?
That’s why I’m always skeptical about measuring AI impact based solely on quantitative metrics.
1 reply →
> writing code has always been the easiest part
Let me provide some possible evidence against this: so many teams are desperate to rewrite their codebase but struggle to actually do so. And when they finally make the leap, it takes them 5x as long as they had hoped. Then sometimes the new code isn't even any easier than the old code.
I personally find writing code to be a huge time suck and I'm very happy that AI helps me save that time.
But again, how much of that time is spent writing code?
I've done rewrites, replatforms, a bunch of times. The actual programming is not the tricky part, but instead (1) picking apart the legacy system, understanding what to build, (2) orchestrating the work to shift transactions from service A to service B without breaking anything.
Teams and especially developers love to think they can skip that phase and just crack on with the programming, invariably what happens is the same as described above: intoxicating velocity followed by a hard stop when you realise you haven't solved problems (1) or (2) above
Most of a software project's lifetime will be spent as a maintenance challenge. i.e. How do we add the 237th feature without adding to the performance problems that already exist. Hence, the desire to rewrite the codebase to incorporate the abstractions of all 236 features.
I don't see AI helping with this. From my experience, it seems like the opposite. It can help you write the code after you've deconstructed the problem yourself and know how to keep it in check.
I reckon a lot of re-writes tend to take "them 5x as long as they had hoped" and "isn't even any easier than the old code" exactly because writing the code wasn't the problem in the first place.
It's business logic, edge cases and other small necessary details that accumulated over time which make the code 'messy'. Once you've integrated all those in the new system, it likely looks equally messy. And discovering and implementing all those extra requirements is probably what took you the longest.
Not to say this applies to all re-writes or that AI tools can't help the process
> “…deciding what to do and what not to do is something else that takes forever…”
If that’s the case, then I should think business owners and office workers should be able to sort that out, lestwise on the “how to automate the boring stuff” front. That repetitive, boring, time consuming, error prone work. Incremental, least work for greatest impact.
The danger is they pull a 1999 Mars Climate Orbiter —level mistake. Or their solution suite grows to big to manage with mounting tech debt.
Also, if you’re software also required some custom domain hardware, then there’s your bottleneck for protectionist business practices.
and also:
* several companies require their tool to have several certifications (SOC 2, ISO 27001 etc etc), how this will work with vibe coded tools?
I see that Software as a Service banked too much on the first S, Software. But really customers want the second S, the Service.
When you sell a service, it's opaque, customer don't really care how it is produced. They want things done for them.
AI isn't killing SaaS, it's shifting it to second S.
Customers don't care how the service is implemented, they care about it's quality, availability, price, etc.
Service providers do care about the first S, software makes servicing so much more scalable. You define the service once and then enable it to happen again and again.
They didnt, dont make the mistake of thinking Saas companies are just software companies. They are Sales companies who happen to sell software. Companies like Dropbox & Atlassian have long been surpassed in Tech but they live only because they continue selling even when demand was hard to get. Their moat is sales & networking and software has to be just good enough. And other part is service, these companies still have one of best costumer service since the start of early 2010s. You can still get refund on Uber quite easily, but if you try doing that at a regular old school company you would require a prayer and couple of business weeks.
Good point, sales is the winning factor in most cases. Why is Microsoft one of the largest software companies? Sales.
> it's quality, availability, price, etc.
Are you sure? Companies still use SharePoint Online, Teams etc.
The F in SharePoint stands for fast
Yes, many don't like Sharepoint, but still they use it. It's the tool they can use.
Customers don't care if Sharepoint uses LLM, they just want to share ideas, files, reports, pages, etc. If LLM makes it easier, great! If some other product makes it easier, great!
It's not about the product it's about the results.
You're proving the point? Sharepoint, teams: availability + price. Every company has microflows, sharepoint and teams are automatically available and part of the price or lower priced than the competition.
1 reply →
Nah it's not that at all. Most of the services are totally fungible and everyone has a short attention span. You need to be in a market which is extremely difficult to disrupt and have a product which people are totally dependent on. And those tend to have a rather large cost to enter unless you were in early.
That 2nd S is sometimes engineered into the product design to maximize vendor lock in, and consulting revenue.
Yes and that is exactly why they are losing. They have hostages not customers.
I think customers sometimes want the third S, Security, which is not (yet) a given in LLM-assisted coding.
I just don't want to pay $50/user/month for an initially open source product that was relicensed and then crippled that the initial group giving something away decided they wanted to make a business of it.
Use the original open source version. They can’t relicense anything they can just use a new license for future versions.
Why not, if it solves your problem?
I think one of the interesting things here is that AI doesn't need to be able build B2B SaaS to kill it. So much of the overhead of B2B SaaS companies is thinking about multitenancy, intergrating with many auth providers and mapping those concepts to the program's user system, juggling 100 features when any given customer only needs 10 of them, creating PLG upsell flows to optimize conversions, instrumenting A/B tests etc...
A given company or enterprise does not have to vibe code all this, they just need to make the 10 features with the SLA they actually care about, directly driven off the systems they care about integrating with. And that new, tight, piece of software ends up being much more fit for purpose with full control of new features given to company deploying it. While this was always the case (buy vs build), AI changes the CapEx/OpEX for the build case.
And in many cases, it's 12 features, with 2 of the features not even existing in the big SaaS.
I'm pretty sure every developer who has dealt with janky workflows in products like Jira has planned out their own version that fits like a glove, "if only I had more time".
If companies wanted to build thier own simple-JIRA they could have built themselves before. I dont think making a kanban board was hard even before AI.
1 reply →
JIRA especially, and I'm always shaking my fist at Atlassian that simple APIs or workflows or reports aren't already included in the tool. I have to pay some other company $10/user/month to get this dumb report your tool should already be able to do?? Insane.
Jira has had free competitors that do at least 75% of what it does since it's inception. You could find a dozen on github that actually look good right now.
In spite of this, Jira is bigger than ever.
Until a given company decides they need access control for their contractors that's different from their employees, etc. etc. etc. - seen it all before with internal often data scientist written applications that they then try to scale out and run into the security nightmare and lack of support internally for developing and taking forward. Usually these things fizzle out when someone leaves and it stops working.
Bingo; the exact same arguments against regular-coding it in-house apply to vibe-coding it in-house.
Pretty much. My employer was looking to cut costs and they were spending ~500k a year on a product that does little more than map entra roles/groups to datasets and integrated with a federated query engine through a plugin. Took a couple days to build a replacement. The product had only a few features we needed.
I've found in the embedded space that people sell lots and lots of products that do everything you could ever want, and the most efficient thing to do is not buy those things and instead find a way to do just the subset of things you care about with your own back-end systems. The upshot of that is that because you're in total control if something goes wrong you can fix it without getting 6 people on a phone call to point fingers at each other.
As niche SaaS provider, I'm trying to avoid succumbing to the same fate. The product I built carefully for years would now be within the reach of a senior dev with a couple focused weeks -- if they knew all the requirements. To avoid being overtaken, I'm working to increase my customer's requirements -- getting them hooked on new reports and features I never had time to build before LLMs could do it for me. This makes it less likely for a competitor to be able to afford to quickly replace me.
At the same time, I have no idea what the cost of LLMs usage will be in the future. So I'm working to ensure the architecture stays clean and maintainable for humans in case this kind of tooling becomes untenable.
1 reply →
there's no shortage of software engineers, if it was so easy for an organization to replace a saas with something built in-house they'd be doing it all the time. In my experience in enterprise consulting implementing a well defined requirement is the easiest part. Getting everyone to agree on the requirement, getting it defined, and stopping it from changing after every demo is the hard part.
Exactly, a lot more focus -- and most importantly specific domain knowledge -- allows the end-user to build exactly what they need, fast.
Sure, and Github is filled with these 10 feature competitors. Some even have active communities.
Didn't seem to kill off the big SaaS players or even weaken them.
I don't think it is killing SaaS. I have definitely had to extend my sales cycle when a potential customer vibe-coded a quick fix for a pain point that might have triggered a sale a few weeks earlier, but eventually the benefit delivered by someone else caring about the software as their entire mission really wins out over a feature here and there.
If you are selling SaaS consider that a vibe-coding customer is validating your feature roadmap with their own time and sweat. It's actually a very positive signal because it demonstrates how badly that product is needed. If they could vibe code a "good enough" version of something to get themselves unstuck for a week, you should be able to iterate on those features and build something even better in short order, except deployed securely and professionally.
Everyone's going to talk about how cool their custom vibe-coded CRM is until they get stuck in a failed migration.
Yeah I have been saying this since the start of vibe coding, Saas companies rely on their sales, who are good enough to sell ther products even in tougher conditions. Software costs for the companies is 100% tax deductible, and they spend a very little on it (Most of times its less than 1% of CapEx). Only reason to optimize this cost is if the Execs of those companies think you can sell the same product.
> Software costs for the companies is 100% tax deductible
Section 174 almost changed that, but was reversed last year.
https://www.goodwinlaw.com/en/insights/publications/2025/07/...
> Everyone's going to talk about how cool their custom vibe-coded CRM is until they get stuck in a failed migration.
Failed/partial/expensive migrations is the name of the game with SaaS as well. Lock-in is the bottom line.
Migrations become much less scary when you truly own your data and can express it in any format you like. SaaS will keep sticking around, especially those that act like white-hat ransomware.
The other thing is bringing in the knowledge about what other customers in the same field want. For business-focused software this can be a boon, customers often can't really envision the solution to their problem, it's like the Henry Ford attributed "If I had asked people what they wanted, they would have said faster horses"
> What they don’t know, though, is that a poorly architected system will fail, eventually. As every senior programmer (eventually) understands, our job is complex because we have to understand the relationships in the real world, the processes involved, and the workflows needed, and representing it in a robust way to create a stable system. AI can’t do that.
I have a strong feeling the future's going to look like this:
Company vibe codes to replace a SaaS.
Little do they know this creates a time bomb: fragile systems where fundamental architectural defects are papered over by humans who knew the underlying dynamics but didn't articulate them well enough during the initial "vibe-architecture," so they're forced to patching the "impedance mismatches" with data entry or with even more vibe coding.
Those humans are eventually laid off, because of course they are. Data quality rapidly deteriorates. Operational mishaps deteriorate relationships with human counterparties. Defects begin to cost thousands to millions.
Suddenly, there's demand: not for SaaS, but for actual service businesses. Consultancies that can parachute in, do actual domain-driven design, and un-vibe that code. They do have a stronger-than-ever pool of out-of-work engineers (many from the failed SaaS companies).
The SaaS companies that survive understand that the first S no longer stands for Software; it stands for Solutions.
You're assuming that code in the future will need to be unvibed. Either the code will be good, or AI will be good enough to unvibe it. That might be awhile in the future but it will happen.
> or AI will be good enough to unvibe it
it already is
just point Claude Code at a Claude Code codebase you forgot about for a few weeks with no plan.md or agents.md or memory.md implemented
configure the session correctly this time and put it in plan mode, it will deal with your whole database schemas, migrate, tie everything to the new models correctly, make a backend deployment script, and fix your UI/UX
2 replies →
It’s about the "Per User Tax"
The panic over SaaS vs AI is simpler than people think. For years, we’ve been paying "Enterprise" prices for tools that are essentially just a UI on top of a database.
I'm a solution architect, and we recently looked at the $30/user/mo price tag for legacy test management tools. It’s insane. Why am I paying a "per user per month tax" for a glorified spreadsheet when I can pay $20 for an AI agent that can build me a custom version in a week?
So, we did exactly that. We used Claude and Cursor to "vibe-code" EZTest. A 100% open source, self hosted alternative that does 90% of what the expensive SaaS tools do, but with zero recurring fees and total data ownership.
The market is crashing because the "Application Layer" has been commoditized. If you can build and own your infrastructure for the cost of a few API calls, the era of renting basic software is over.
We aren't just building a tool; we're proving that the "SaaS Tax" is now optional.
Literally everything is a UI atop a database. Writing the data models, crafting the right uI and the right flows is non trivial. It requires iteration and that refinement is what customers pay for. The reality with SaaS stocks is they dont make anything critical. eg Take any large consumer tech company and they dont use Klaviyo. These companies build their own stack and one thats intimately connected with the data lake. It works for tech companies since they hve good data and it doesng work for non-tech companies because a good fraction of their data is trash.
That’s great. The overhead is that you have to host it and maintain the environment and app. If there is downtime that’s now your responsibility.
If that’s better than $x/month to be someone else’s problem then it’s a win.
Is there a middle ground that hasn't emerged yet?
Someone to host and manage your SLop-As-A-Serivce (SLAAS) for a price point in the middle.
That feels like a business.
1 reply →
> How to keep asking customers for renewal, when every customer feels they can get something better built with vibe-coded AI products?
Wrong take. You don't need to build something better, you only need something good enough that matches what you actually need. Whether you build it or not and ditch the SaaS is more of an economic calculus.
Also, this isn't much about ditching the likes of Jira not even mentioning open source jira clones exists from decades.
This is more of ditching the kind of extremely-expensive-license that traps your own company and raises the price 5/10% every year. Like industrial ERP or CRM products that also require dedicated developers anyway and you spend hundreds of thousands if not millions for them. Very common, e.g. for inventory or warehouse management.
For this kind of software, and more, it makes sense to consider in-housing, especially when building prototypes with a handful of capable developers with AI can let you experiment.
I think that in the next decade the SaaS that will survive will be the evergreen office suite/teams, because you just won't get people out of powerpoint/excel/outlook, and it's cheap enough and products for which the moat is mostly tied to bureaucratic/legal issues (e.g. payrolls) and you just can't keep up with it.
Having participated in the build of an inventory system / system of record for a large national retail company, I can't see vibe coding helping anything more than the prototyping in the discovery / requirements gathering parts of the process.
The sheer volume of data, the need for real time consistency in store locations, yada yada means that bad early decisions bite hard down the road.
Lots of drudge work can be assisted by AI, especially if you need to do things like in ingest excel sheets or spit out reports, but I would run far away from anything vibe coded as hard as possible.
The example I made about inventory wasn't random.
One of my clients spends 500k+ on XXX licensing per year (for a 200M revenue company that's not peanuts), and on top of that has to employ 12 full time XXX developers (that command high figures just for their expertise on that software while providing very little productivity) and every single feature takes months to develop anyway. Talking about stuff like adding few fields to a csv output.
So the total cost of XXX is in the 2M/year range, and it keeps ballooning.
My (4 men) team already takes care of the entire warehouse management process except inventory, the only thing that XXX provides, we literally handle everything: picking, manufacturing, packaging, shipping phase and many others.
In any case, nobody has mentioned vibe coding.
I stated that a handful of good engineers with the aid of AI in a couple of months can provide a working prototype to evaluate. In our case it's about extending our software that already does everything, except inventory management.
When you spend 2M/year on a software (1% of your revenue), growing every year by 100/150k it makes sense to experiment building a solution in house.
4 replies →
Its funny you mention excel, I see vibe coding in the business sense right now being a gateway to replace all of the ad hoc uses of excel. We've basically leveled up the quality of the software you can build before buying a SaaS product or a hiring an in house engineer.
4 replies →
"For example, to create a data visualization I won’t seek any SaaS. I’ll just code one myself using many of the popular vibe coding tools (my team actually did that and it’s vastly more flexible than what we’d get off-the-shelf)."
That maybe doable in your 10-people startup, Namanyay. Try doing it in a larger organisation with layers upon layers of firewalls, databases, authentication systems and not the least importantly - management. Not to mention the vastly different audience, both in size and interest. Your own experience is not the experience of everyone else.
also, who pays for "a data visualization" SaaS?
I guess they mean BI, but for a company of any scale, they aren't paying for a chart, they're paying for a permissions system, query caching, a modeling layer, scheduling, export to excel, etc.
Stand alone BI tools are going to struggle, but not because they can easily be vibe coded. It'll be because data platforms have BI built-in. Snowflake is starting down this direction and we're (https://www.definite.app/) trying to beat them to it.
Definite looks pretty interesting!
1 reply →
The term "B2B SaaS" is doing a lot of heavy lifting here and I think conflates two different things:
(1) Business model: hosted software you pay monthly for (vs self-hosted/one-time purchase)
(2) "Glue" products: tools like Monday.com that primarily provide synergy between data sources and workflows
The article is really about (2) - and yes, those are vulnerable to vibe-coding. If your product's core value is "we connect X to Y and show you a dashboard," that's now a weekend project.
But there's a huge category of SaaS where the value is in the product itself, not the integration layer. Take Excalidraw - fits the SaaS model, but try vibe-coding a collaborative whiteboard with real-time sync, proper data persistence, conflict resolution, export formats, etc. The hard problems aren't "connect API A to API B."
Or PostHog - sure you could vibe-code some analytics tracking, but building reliable event ingestion at scale, session replay, feature flags with proper rollout controls? That's years of engineering.
The "vibecodeable" SaaS products were always somewhat commoditized - AI just accelerated the timeline. The ones solving genuinely hard technical problems seem a lot safer to me.
I don't really agree with this.
Simple CRUD app sure, but we're nowhere near being able to vibe code even a relatively low-complexity enterprise SaaS product.
If it's got customer data in it and/or you're making important business decisions based on it, you really need your system to be accurate and secure. My experience is the people who procure enterprise software know this and tend to care a lot about it. They often have legal and contractual obligations around that.
In the 1990s there were people who thought OOP with point and click tools like FoxPro and Delphi would make it so easy to create software that everything could be built in-house without expert programmers. The invention of SQL was supposed to eliminate roles like Report Writer and Data Analyst because now business people could just write their own queries "in English" and get back answers.
> In the 1990s there were people who thought OOP with point and click tools like FoxPro and Delphi would make it so easy to create software that everything could be built in-house without expert programmers. The invention of SQL was supposed to eliminate roles like Report Writer and Data Analyst because now business people could just write their own queries "in English" and get back answers.
And yet, precisely that happened in the end, just not with the tools envisioned. Excel, VBA and, where you had one knowledgeable employee, MS Access makes for incredibly powerful and incredibly hard to maintain "shadow IT" - and made even more difficult when someone sneaked in a password, because that takes a bit of an effort to remove [1], knowledge that is easy for us today to find, but not when I was young.
Also, back in the IE6 era, there was a lot of point-and-click created web interfaces... just that it wasn't HTML5 or even HTML. It was an <object> tag loading some ActiveX written by some intern in VB6, or Java, or Flash. I sort of miss that era but also, it was a damn security nightmare. Flash with its constant stream of security vulnerabilities was ripe for exploits, but at least it didn't run native code with full user privileges by design. I'm not kidding, theoretically you could go and import/use functions from any system DLL up to and including Kernel32. OLE/OCX, ActiveX... a design way ahead of its time.
[1] https://stackoverflow.com/questions/272503/removing-the-pass...
Software got easier to develop, but we just came up with more problems to solve with software.
The new tools didn't shrink demand for COTS enterprise software - it grew massively since the 90s!
2 replies →
I think opensource is a good analogue here. For many SaaS products, you don't even need to vibecode anything - there is already a reasonable OSS alternative. Yet people still pay for the SaaS. They want support, maintainability, security, edemifcation, a throat to choke, regulation and domain expertise, etc.
I do think like this HN post (https://news.ycombinator.com/item?id=46847690) is a good example of where a custom more domain specific solution makes a lot more sense that dropping in an off-the-shelf ERP. Still though, I think the bakery would prefer to buy the bakery-ERP than build it but vibecoding does reduce the barrier to entry so we might see more competition and share taking from incumbents by domain-specialized new entrants.
Lots of companies buy saas, and then spend years customizing and effectively building what they thought they bought. And for big companies, it is costly - a few tens of millions for saas licenses, and maybe around 50-100m for system integrators leaching on the enterprise, and doing the integrations and customizations, usually dancing around the data model, api surface limitations of each of the saas tools they want to wire together.
I dont think going back to having own developers, owning the code is going to be a bad financial propositions for such companies. My company is now one month into trying this out and so far, so good. We kicked our outsystems addiction and are just went live with a react rewrite - and are well into rewriting an expensive to run document management system which we were at the same time under-utilizing and abusing. Our product people are loving it since for the first time in ages we dont need to tell them "well that would be real hard, considering we have salesforce crap underneath and it just doesnt do this or that well".
My thoughts are sort of similar. AI/LLMs have allowed many of the typical SaaS customers to think that maybe they could do at least some of this work themselves, and get better results.
For most companies this was always true, but LLMs have given them the confidence to actual start writing more software in house. The SAPs of the world have nothing to fear, companies aren't going to vibe code a CRM, but they are going to be able to more easily write integrations. At a previous job we frequently had bills for €10.000 for small integrations into our ERP, but once we figured out the API and gained more confidence in our abilities, all those integrations got pulled in house.
If your SaaS platform provides actual benefits, then you don't need to worry. If your business in writing integrations for other companies into platforms you don't own, yes, AI is going to hurt your business.
This should have happened regardless of AI though. The idea that companies (over a certain size, e.g. ~20 people) doesn't have a least on developer employed, regardless of industry always seemed like a missed business opportunity. We wrote so many tools for sales, warehousing, customer service and accounting and it's hard to imagine the business functioning without those tools. I might have spend two weeks writing a tool, but if it saves sales 20 hours a week punching in orders, we get a positive return in a month.
That sounds great. What SaaS are you replacing?
I'd rather not name names - but one of the major, popular ones.
We contracted a lot of usage, and are using it literally like a S3 bucket with a malware scanner attached to it, and ignoring the dozens and dozens of document management capabilities it has - that we don't need. (Because really, we only ever needed a S3 bucket with a virus scanner...). This alone will allow us not to renew that contract, and save, maybe, around 2M per year.
Sure, we will have to have our own API that will require support and what not, but... we already HAD to have our own API that requires support and what not, since we have a bunch of legacy document management platforms running in various countries, and we anyway have to operate an abstraction and a router.
I am sure ours might not be the most typical case, but there will be savings, and since the economy is what it is, my bosses are telling me to go for every saving I can find, and thats one of them.
(I'd not try to re-write an ERP system, for instance, or a CRM. But a lot of smaller things where we pay a substantial premium? Sure - we will try.).
Maybe it's mostly from AI, maybe it's mostly general economic cutbacks. I also feel like these "wrapper" style SaaS products are the first ones companies are dropping when they are looking to cut costs, and I think a lot of companies are looking to cut costs. I do agree with the overall conclusion either way, that System of Record products/companies are the most likely to survive. There are a lot of SaaS companies with questionable long-term businesses who are getting hit, but that was bound to happen.
The stocks of a lot of these SaaS companies were priced on the expectation that they could become the next IBM: become entrenched with the customer and then hike prices until their eyes bleed.
A lot of companies have been too smart for that, and a lot of SaaS offerings are too small to be truly entrenched. Arguably the investment horizon is too short (IBM took decades getting to that point).
The only real vendors who managed to become the next IBM are the cloud providers.
I think it's a combination of budgeting, upward price pressure from the SaaS companies themselves, plus bringing things in-house through vibe coding, but there's another factor that I think is harming existing SaaS products. Many of them are becoming legacy solutions with AI bolted on top so they don't really feel that effective or next-level. The underlying tech might even be a generation older too - but the SaaS value-add is providing support, scaling, etc to maintain whatever some old tech that's still a requirement. At some point someone looks at all of these interconnected systems and just says 'start over'.
Vibe coding might not be supplanting all SaaS solutions but it's definitely shaking out "last-gen" solutions.
System of Records especially for boring industries is the way to go. What kind of wrapper SaaS are you seeing getting dropped?
Analytical systems. I see a lot of add-on services that will add intelligence/analytics/etc and companies try them out to solve some issue they have and bounce off them frequently due to growing costs. I can only assume as mentioned that over time these are also easier for companies to in-house vibe-code as well, I just haven't seen a ton of that yet, but people are definitely trying which still shrinks the available pie.
This isn't happening. The past six months has been rough on public B2B SaaS valuations, but the impact is a lot wider than just B2B SaaS (its all non-S&P10 software), and valuations are just vibes in the end. Most of these companies are, financially, doing pretty well; seeing key metric growth, including revenue and profit. This makes sense: AI does not fundamentally change the bargain SaaS brought to the table, that companies would rather pay someone to solve their problems than solve them themselves. However, the stock market doesn't care about this. The stock market doesn't care about anything; it behaves irrationally and non-sensically, and trying to derive any sense of how stable, strong, or successful a company is from stock market valuation is like using lines of code to claim that a software project is really good.
>that companies would rather pay someone to solve their problems than solve them themselves.
Are they not able to just engage AI to solve those problems now? E.g. this morning I saw an app that did something interesting to me for $20 a month. 20 minutes in Gemini and I had a functional app that replicated the behavior. SaaS are more complex but give me a small team and a couple months and we could replicate most any of them.
No one is replacing Jira or Salesforce with an internally-AI'd analogue.
1 reply →
Equity markets both private and public are mangled today, for a wide range of reasons which I wont get into.
Financial performance e.g. revenue is what counts right now as any hard-evidence.
This feels a lot like the old RPA hype cycle to me — more sales narrative than structural change.
Most companies are not going to replace stable SaaS with a pile of AI-generated internal tools. They don’t want the maintenance or the risk.
If there’s a real B2B game changer, it’s Microsoft.
The day Excel gets a serious, domain-aware AI that can actually model workflows, clean data, and automate logic properly, half of these “build vs buy” debates disappear. People will just solve problems where they already work.
Excel has always been the real business platform. AI will just double down on that, not kill SaaS.
> The day Excel gets a serious, domain-aware AI that can actually model workflows, clean data, and automate logic properly, half of these “build vs buy” debates disappear. People will just solve problems where they already work
Best they can do is more adware in windows. Sorry.
"The SaaS model was built on a simple premise: we build it once, you pay forever."
I've never seen a SaaS product that fits this description. There are always things to do. Libraries to upgrade, performance bottlenecks to diddle around with, an endless stream of nonsense feature requests from people at the customer who never actually use the product, fun experiments your developers want to try out, and so on.
The hard part in SaaS is to delete code, and that's what you should do, at least some of the time. Either through simplifications, or just outright erasing functionality that very few if any of your customers rely on.
What you should not do is let your customers grow the liability that is code in your production environment, unless your entire product set is designed to handle things like this, e.g. the business models of Salesforce and SAP.
It's not and I really doubt it will, for true SaaS platforms. A desktop .gif recorder (frequent example I've read about) is not a SaaS, even if you charge monthly for it.
Let's put an example an exception-tracking SaaS (Sentry, Rollbar). How do the economics of paying a few hundred bucks per month compare vs. allocating engineering resources to an in-house tracker? Think development time, infra investment, tokens, iteration, uptime, etc. And the opportunity cost of focusing on your original business instead.
One would quickly find out that the domain being replaced is far more complex and data-intensive than estimated.
There are many cases where the company might only use a fraction of the features (and therefore complexity) of the SaaS and so only need to develop and maintain those features they actually need. That's when ditching the SaaS can make sense if you can easily develop/maintain what you specifically need on your own with AI assistance.
Even if they use it less, if you combine all of the Saas products used by a company, thats a tiny fraction of the overall CapEx. And this cost is tax deductible, so there is no reason to optimise it unless Execs are really penny pinching, but at that point that company isn't worth selling to anyway.
What the authors of this kind of doomer-type articles do not realize is that B2B software companies have the data that their customers pay for, and they also have access to the AI tools themselves, meaning they can accelerate in adding new features to their products, making them more competitive.
It's a fallacy to consider the bad performance of software stocks as a definite sign that AI is going to replace them. One needs to factor financials into the equation to explain a downtrend. Take Figma for example, spending 109 mil on AWS bills, cutting through their margins. Investors know that such costs do not simply go down due to the vendor lock-in companies experience when using cloud services.
Claude Code is good, but definitely far away from being able to vibe code Figma.
I think the corrective to this is that many of these incumbents will fail to re-conceive their product stack from a user-centric perspective, and as a result they will be reduce to just a dumb data layer which is easily swappable.
Sure, they could do that, but the cultural change required is an order of magnitude harder than just sticking an agent on top of their source-of-truth and believing that the problem is solved.
Maybe it works for areas where the application is a relatively self-contained island of productivity. Figma is somewhere that a designer spends a lot of their day, so it's going to be less vulnerable, but most pieces of softare fit into broader workflows. So for Figma the disruptor is less likely to be "AI-powered designer" and more "AI-powered web builder" - e.g. Lovable or even Claude Code itself that just generates great designs.
> can accelerate in adding new features to their products
If anything AI helps companies escape the "feature wheel" that is used to justify exorbitant costs, while providing debatable (and often even negative) utility to the end user.
Keep in mind, Excel '98 would still probably be overkill for 85% of people's needs in 2025.
What companies thought was adding more utility, was actually just continually stacking costs in front of getting to that core functionality.
AI replaces the core functionality, and the "feature" scheme collapses...
“Overkill for 85% of people” is a vibes statistic, not an argument.
Also you're generalizing some things to the whole sector like every software provider now is useless and the new features they add are not bringing any values. How can you make such generic statements about a sector that is so diverse?
I start to suspect some of you are just private-equity-sponsored accounts trying to push Anthropic propaganda over the Internet.
2 replies →
Claude Code doesn’t need to vibe code Figma. It _is_ Figma.
Take a couple screenshots of your legacy app, write a short paragraph describing it, and the web tool will give you a self-contained HTML file that’s a fully interactive mock-up.
But it’s still a mock-up. So software dev in general will be fine, it will evolve. But unless the AI companies run out of money and it turns out the $20/mo plan actually costs $1000/mo without VC subsidies, Figma is cooked.
so many conditions need to be met to have Figma "cooked". Plus you forgot that the underlying LLMs need to get exponentially better, hard to do that without infinite VC money.
The software industry moat is disappearing. What will happen is 1000 (10,000?) competitors putting price pressure on every saas app.
Software will cease to be a winner take all and be a very long tail distribution!
Where are all these 1000 competitors coming from?
You need to have tremendous agency/will to start competing with a public company. Plus, you need to have a lot of distribution channels, competing with sales people that do this for a living since ages, and marketing budgets that are higher than your annual Claude spending.
Regular people do not press 10 clicks daily to track their calories, and you're saying that they will start competing with Salesforce and the like?
Focus is a currency and you have a limited amount of it, if all SaaS is built internally, teams would go bankrupt. There's likely always going to be a band of experts focused on solving a problem and everyone pays them to solve it for them, because they do it better and can handle the hassle of maintaining it.
AI isn't killing SaaS exactly, but instead of selling UIs, SaaS companies are going to have to focus on infrastructure and data. You have to host stuff somewhere, so there's an inescapable cost and transaction that has to take place. If businesses can pay one bill for infra + data management and get nice apps and stuff on top of that (without being locked in), that makes more sense than trying to roll stuff together even if you have a platform team.
> AI is killing B2B SaaS
Anecdata sample size of one, but this is not my experience at all. My business has only continued to grow over the past couple years, and I don't think I've had a single customer mention AI to me at all (over the phone or email).
It's the opposite IMHO. AI is enabling a lot more B2B SaaS. There are a lot of companies that are running on outdated software. Especially in manufacturing and industry. They've had decades of experience with very expensive IT projects where cost got out of hand because things just are very complicated in the real world.
There are many millions of companies that are going to be re-examining if they can do better in the next years. The work will still be very complicated but with the help of AI, small IT shops might just deliver enough value to be worth the trouble.
The notion of e.g. busy floor plant/logistics managers vibe coding this themselves is silly. 1) they don't have the time; these people are super busy 2) they lack the ability. 3) they'll want it done properly 4) their employers won't skip all the certifications, iso stuff, and what not.
Companies invest in SAAS software if it delivers some kind of revenue/profit benefits. If it's too expensive/complex, it can't do that. AI tools lower the cost of SAAS solutions. So the totally addressable market grows. Companies will want to maximize their ROI though. So, they'll do the usual and engage software companies and integrators to help them do this. They'll expect to pay less for more. And there will be lots of haggling around that topic. But there's an enormous amount of companies that are quite far behind on getting their operations into this century in terms of IT already. There are going to be early adopters looking for early successes here that will put pressure on their competitors if they are successful.
I'm running a small company in this space. We're seeing a lot of opportunities right now. And AI is making my work massively easier already. All those complex ERP integrations just became an order of magnitude easier to do with a small team. They are still hard though. Forget about vibe coding that. You need a plan.
The framing is a bit dramatic but the underlying shift is real. What AI actually kills is the "wrap an API in a UI" SaaS model. If the value is just presenting data nicely or doing simple transformations, an agent can replace that.
What survives: products with proprietary data, strong network effects, or deep domain expertise baked into the workflow. The moat moves from "we built a UI" to "we understand this problem better than anyone."
I run 4 side projects and the ones getting traction aren't the ones with the fanciest AI features - they're the ones solving specific problems people have repeatedly (meal planning, meeting search). The AI is the engine, not the product.
The real risk for B2B SaaS isn't that AI replaces your product - it's that your customers can now build a "good enough" internal version in a weekend with Claude Code.
why does half of your comment read like it was written by AI?
Fair callout. I've been writing too many product descriptions lately and it's leaking into how I write comments. The actual point was simpler than I made it sound: AI kills "UI wrapper" SaaS, but products with real domain knowledge survive. My side projects taught me that - the ones getting users aren't the technically fancy ones, they're the ones solving boring specific problems.
It’s not as far-fetched as people think. I see so many comments here doubting you can vibe code a full CRM or e-commerce SaaS, but a skilled AI-assisted programmer absolutely can, especially if they're aware of strong open-source alternatives already out there.
For Salesforce-like CRM, there's Twenty[0], a good-enough alternative. For Shopify-style e-commerce, Medusa[1] is a headless commerce platform.
The real power comes from using AI to study how these projects implement specific features (payments, inventory, customer dashboards, etc.) and adapt them to your stack. AI excels at finding the "seams" (those connection points where a feature ties into the tech stack) and grasping the full implementation. The trick is knowing precisely where the feature lives in the code (files, functions, modules), because AIs often miss scattered pieces otherwise. That's what I'm building at opensource.builders[2]: turning OSS repos into a modular cookbook with structured "skills" that point to exact details for reliable remixing and porting.
SaaS companies are forever beholden to raising their market cap, even in solved spaces like cart, payment processing, and CRMs. Most businesses run on CRUD apps anyway, and if your core app exposes an API, you can build any customization you need on top of it. People here discounting how valuable it is for a business to have the software that runs their business on a tech stack they understand and something they truly own.
[0] https://github.com/twentyhq/twenty
[1] https://github.com/medusajs/medusa
[2] https://opensource.builders
Just because it's possible to build equivalent software by vibe coding doesn't necessarily mean that companies will stop using SaaS. There are multiple reasons why...
First of all, many big companies pay a fortune to use inferior SaaS solutions instead superior Open Source solutions; possibly because one of their CTO may have received kickbacks or promises of a lucrative job at the SaaS provider as a consequence of this deal. There are a lot of politics going on behind the scenes when it comes to procurement.
Execs at big corporations are often looking for plausible ways to spend investors' money in a way that they can capture some of it for themselves. If they choose open source or they choose cheap vibe coded solutions; there is not much money changing hands. No opportunities for insiders to covertly monetize.
And then there are a lot of security implications to using a complex vibe coded app. The AI won't be able to identify the vulnerability in any decent sized codebase unless you know what you want it to look for.
Lets break this down. There is very little in newness in what Anthropic announced. Claude had skills for a long time. They have added one more layer of abstraction and called it plugins. This mainly comes with a set of integrations.
Thats the pitch.
But, what are Claude plugins?
Plugins=Commands+Skills+Integrations.
Commands are specific to Claude code. But commands and skills are nothing but prompts at their basest level.
So what is the main differentiator?
Integrations.
But what are you integrating with?
SaaS companies.
And what is the stock market doing?
Dumping SaaS stocks.
How do they think Claude cowork will work without the integrations. Without the system of records.
If anything, these SaaS products have become more important. If I was a trading guy, I would go to the github of claude plugins, see the default integrations and buy the stock of those companies.
Lot of places that I see AI disrupting - I'm not buying that SaaS is going to be a significant one.
Reading through the article:
> They were paying $30,000 to a popular tool3
Couple things we needed to understand here:
If it's a technology company of > 1000 employees - then $30,000 month doesn't even get Finance's attention. And there is next to zero chance that anyone is going to vibe-code, deploy, support and run anything in a 1000 person+ company for $30,000 a month. SaaS wins hands down.
Any product/service that people care about comes with a pager rotation - which is 6-7 employees making > $200k/year. If you can offload that responsibility to a SaaS for < $1mmm/year - done deal.
Yeh but in a company of 100 employees for software of 30k a year, it's more than worth it to take your standard 50k (GBP) dev and have them replaced it. It's a one time cost, and the support time will certainly be less than 50% of their time every year so it saves money.
There are many companies that operate like this all over the world. Outside of the hyper-growth tech VC world cutting costs is a very real target and given how cheap Devs are outside of America it's almost always worth it.
$30k/year? For 100 Employees. So - $25/seat?
I can't imagine it would ever be worth, under any scenario, trying to write/build/support any $25/seat SaaS software for any company I've worked at in 25+ years.
Another thing to keep in mind - very little of the cost of a SaaS license is the time it takes to build the software. Security, Support, Maintenance, Administration, backups/restores, testing/auditing said backups/restores, etc, etc.. and then x-training new SREs on how to support/manage this software, ...
Even as someone who spend 10+ hours a day churning out endless LLM applications, products, architectures from my myriad of Cursor/Codex/CC interfaces and agents - I'm dubious that LLMs will ever eat into SaaS revenue.
I'm sure (lots of) people will try - and then 1-2 years in someone will look at the pain, and just pull the ripcord.
I don't see that happening because companies need to concentrate on their differentiators. Is your enterprise vibe coding its own SaaS? Who's taking care of it?
> Is your enterprise vibe coding its own SaaS?
Yes, a lot.
> Who's taking care of it?
It's not hard.
We wouldn't do it for tools that are purpose made and have sane pricing in the market place. We do it for stuff that would traditionally go on a 'platform' like Salesforce or something that requires a lot of customization to begin with. It's so much easier to just roll your own than even just going through the procurement process of those kinds of tools much less the integration and change process (hiring consultants, etc). I'm not hands on with it, but I know our small group of AI are helping us eliminate $5m recurring annual spend this year and that's directly impacting the topic article. I won't be surprised if at some point we replace our more sticky ERP software or use this leverage to negotiate prices that are sane. Businesses have been gouged by enterprise software long enough.
Am I to understand your company wrote a CRM? What other applications did you replace? What company is this?
1 reply →
Imagine working at a company who has it own Figma, Docker etc... Thats a recipe for disaster.
And you'd have to relearn everything every time you changed companies.
AI
It's hard to tell when people are joking.
Boy that "st" ligature in the subheading font is eye-catching, to the point of distraction.
There's a lot of it in the article. Very distracting, I just didn't know what to call it (I searched for 'tail' before posting my own comment).
Indeed totally unbearable to even read the article
Agreed. I thought it was weird on the ct ligature on something like the second sub heading and the next one had a st ligature. I stopped reading the article and was just scanning for st ligatures in the text itself (there are none) and then realised the main font was a less legible serif font and the headings that were bigger and shorter were sans serif which makes the ct and st ligatures stand out even more.
I didn't read any more of the article after that, and the primary reason was just this weird font choice.
I stopped reading the article because of it.
It really isn't. Nobody's migrating their payroll vendor to some vibecoded app. I agree with Jensen Huang that this reaction doesn't make sense at all.
My own prediction is that reliable vibecoding will be additive. It's a new capability that will help high-agency people do things faster or answer questions they couldn't easily answer before. Need to spin up a big custom Monte Carlo calculation and want a simple UI to control and configure it? You can just throw that together now. Need to get a draft budget allocation for a big set of projects calibrated against a set of conflicting constraints? Let an agent crank at it for a couple hours, then review and refine manually -- or just toss it out if you don't like it.
But building, running, and maintaining production-grade services or apps that the company relies on for its basic functioning? You're not just paying the SaaS vendor for having built the product -- you're paying them to maintain it, run it, and respond to issues promptly. You're also paying them to keep building it and improving it over time. To be clear, I think there are certainly cases where the rise of "coding AGI" is going to lead companies to build some services internally versus buying from a vendor, but I predict these will be highly custom and bespoke services that are too tailored for a specific corporation to make sense for a third-party vendor to try to sell.
Here is the list of evidence the author gives for why AI is the reason software company stocks are down:
but it is down 25% in days. if not contagion that ai is going to kill saas, then what?
It seems like 'the market' is making this bet. I'm not deep into financial reports or whatever. But what I'm seeing from the tech side, this is not at all true.
If anything B2B SaaS is growing with AI, and it hasn't even begun, the biggest AI markets right now are personal. The B2B market is up for grabs for sure, 0%-1% of niches have an LLM product right now. But traditional SaaS has a huge advantage, they have reams of industry specific data, and they have the customers, sure they will have competition, but they are the incumbents.
If I had any money I'd buy the dip
As a founder, there is another angle here that is worth mentioning. Not only does AI B2B SaaS allow insourcing, it also allows there to be 10x (imaginary number) the number of companies building SaaS for the same use case. What we see in healthcare or finance for example is executive fatigue from demos, in many cases mostly vibe coded frontend UIs that entrepreneurs are using to test the market. This creates friction for businesses / SaaS companies that are unable to show how their solution is unique, well built or has a clear moat over the many others they have seen.
Bang on. I mentioned this as well. Mature SaaS companies are also expanding into each others domains. Notion is now doing email for example.
While the author is wildly overstating things, I do think AI is striking at the heart of the SaaS problem, which is the business model of "pay us $10-100+ per employee per month in perpetuity or we will hold all your data and your company's operations hostage". There is always going to be value in good software, but it is shitty vendors relying on the lock-in effect that are in danger. And good riddance.
The other issue is valuations - B2B SaaS stocks have never been rooted in reality, and the 100+ P/E ratios were always going to come down to earth at some point.
Agree on the valuations. Most have come down and many have overcorrected imho.
As expensive as some of these software seem in terms of cost per seat, most of the subscription contract rarely exceed a few hundred thousand / year if even $1mm, which is drop in a bucket for many companies. (vs running on-prem servers, having staff to support them)
You'd think Atlassian would be printing money given everybody under the sun is using them, but they only make $5B in annual revenue.
I've worked at fortune 50 companies for a while and custom enterprise software is still alive and well for things that are too business specific to buy off the product for. But they're not going to be in a rush to create their own Workday, Salesforce, Jira, Figma, SAP, etc.
Sure its fun to (vibe) code some internal version for a SaaS, but maintain it month after month? Maintain SLAs, etc? That's not fun.
Vibe coding gives you that dopamine hit of creation, but does the internal dev really want to deal with the care and feeding of the random shitty timesheet app they created?
Do they want to take on the work of integrating random backend systems that timesheet system needs to talk to? Do they want to get called at 3AM when it's down?
Even AI assisted, living year after year with production systems is hard.
> we have to understand the relationships in the real world, the processes involved, and the workflows needed, and representing it in a robust way to create a stable system. AI can’t do that.
That is because AI is living in our world, instead of the opposite where we live in AI's world.
Case in point: maybe the AI hallucinated a class method that never existed in our world yet, but perhaps in the AI led processes and workflows it would be written to better fit into the smooth gradient decent those same top parameters' scores.
I think there may be other factors killing SaaS, particularly data sovereignty.
"According to IDC’s Future Enterprise Resiliency and Spending Survey from June 2025, 45% of all organizations and 56% of “digital natives” cited data sovereignty and potential cloud changes as their greatest concern for 2026."
https://www.veeam.com/blog/saas-data-sovereignty-microsoft-3...
“Killing” is a bit strong, but is there a world where folks just vibe code solutions that they would have bought previously? Absolutely and and I think that world is here now.
I’ve seen many startups recently were it was like “guys I could vibe code your ‘product’ in the afternoon.” Yes someone needs to look after it etc, but the bar on where companies buy vs build is getting much, much higher.
(Insert rant from dev teams about the code sucks, who will maintain it, etc). Yes all valid points, but things are changing regardless of if folks like it or not.
Counterpoint, all your customer's data just got stolen by teenagers in the 3rd world and is available for purchase on the dark web. Was it worth saving $5k a month?
A lot of startups/small businesses are like "with AI we can build more than ever". The problem is so can everyone else and capitalism rewards scarcity not value. The bar for startups and small software business has risen quite substantially. I know we are avoiding buying software now where I work if possible unless we previously committed to it (contracts).
Saas companies will survive for the same reason they do today. The operational overhead of any sufficiently complicated piece of software is too much, even more so if it's vibe coded.
The bus factor is gonna be pretty high if your enterprise relies on an internal tool that some guy at your company vibe coded at some point.
Bus factor would be 0 because even he wont be able to debug it.
This was true pre-AI, but now, the bus factor is actually way lower in any software than it was before:
- Hey Claude, what is the project in XXXXX/ about and how does it work ? What should be improved there ?
1 reply →
One problem with centrally produced and distributed software is that a small subset of users demanding certain features results in feature bloat for everyone. Costs for all features are shared by all users.
Probably one way SaaS companies will adapt is to break up their offerings into more modular low cost components. While many customers will end up paying less, the addressable market will probably increase because of the new low cost options.
> Costs for all features are shared by all users.
To a degree but most enterprise focused software usually has differential pricing. Often that pricing isn't public so different companies get different quotes.
The framing of 'vibe coding replaces SaaS' misses the more interesting shift: the value SaaS provided was never really the software — it was workflow automation. Software was just the best delivery mechanism we had.
What's changing is that agents + APIs are becoming a better delivery mechanism for many workflows than a UI you manually operate. A company paying $50k/year for a marketing analytics dashboard doesn't actually want a dashboard — they want answers about what's working. An LLM with API access to their data sources often delivers that faster than navigating someone else's opinionated interface.
The SaaS most at risk isn't infrastructure (Stripe, Twilio) or systems of record (Salesforce, Workday). It's the 'pretty UI on top of data you already own' tier — analytics, reporting, simple automation, basic CRM. That's where the compression happens. The products that survive will be the ones that become the system of record, or that offer value AI genuinely can't replicate (regulatory compliance, deep integrations with legacy systems, etc).
I didn't realize B2B SaaS products were in freefall like his numbers suggest. I'm not convinced customers are leaving to vibe code their own products but I do believe we're seeing a major shift in the market, pushed by the sudden relative ease of coding. There are a lot of B2B SaaS products which are outdated and I wouldn't be surprised if they're supplanted by much faster competition
Yup it's definitely not because _customers_ are coding solutions, but the trend and motivation seems to come from the fact that customers are realizing there's something else possible except being tied into expensive recurring yearly subscriptions.
I was surprised when I saw the numbers from Bloomberg myself as well!
The keyword in "Software as a Service" is not Software; it's the Service.
In the early days, the tagline for Salesforce is "No Software". It's secret recipe is this: your sales team only need a browser and a credit card, to get the service. No software installation needed. Even if you have a genius can code something equivalent, it will never be a "service". That genius is not going to support it, not going to add storage for you, not going to restore an accidentally deleted record for you. That takes an army to deliver. It is a service.
Of course, Marc Benioff kind of shot himself in the foot by trying to get ahead of the AI curve... and gutted their customer service division. If the service is delivered by AI agents, what is the selling point again over other AI agents? They have debased their key strengths and are getting punished for it.
I used to be a big advocate for Salesforce in my organization. And it was really great .. allowing us to deliver new functionality without the usual IT procurement bureaucracy.
Now with cloud maturity and Vibe coders who will get better and cheaper, I think it's possible to replace all the features we use on Salesforce at a fraction of the cost of our Salesforce licensing cost.
counterpoint: why would sales people want to use a different crm each time they job hop every 2-3 years?
There is no evidence presented that internally "vibe coded" products are the reason hubspot et al are struggling right now. If anything, the fact that the divergence from the broader index starts in April of last year (well before the current vibe coding moment got going) is evidence that this is something else.
Vibe coding seems to be the iPhone camera to DSLR moment for programming.
- No professional used an iPhone for years. Most don’t today.
- Professional scoffed at it as a toy
- The toy shifted the balance of volume through everyday enablement of amateurs to a degree that professional were right, but now in a severely lopsided terrain.
The value ends up in the most engaged paradigm, rather than the most perfect one.
I just don't buy it.
Most people who've been in a business SaaS environment know that writing the software is relatively the easy part aside from in very difficult technical domains. The sales cycle + renewals and solution engineering for businesses is the majority of the work, and that's going nowhere.
Great insights Namanyay!
I feel the same way about needing to build platforms that enable companies to build on top. That's why i'm building adventureflow.ai to enable energy companies to build their agentic workflows directly on top of my platform.
The expectations have completely changed now, customers expect more software for less, because it is 5-10x cheaper now or faster to build and there's no changing that. Tools like Opus 4.5 have completely disrupted the software marketplace!
Salesforces market cap today is 180,35B. That, not the change, is the market's judgement of how valuable salesforce is, despite any risk from AI. If salesforce stock drops with a new AI model that doesn't mean that investors think salesforce is worthless. It means a slight risk to their business is getting priced in.
Not sure about that, however agents in low code tools are certainly taking over old school integrations.
Nice, what kind of agents and integrations are you seeing being used?
Platforms like Boomi, Workato, Optimizely Opal,
Wow, the tails on the lowercase letter 't' throughout that article are distracting.
I am kind of hoping that AI will kill the startup grab for money TBH. Too many wannabe CEOs I've met in the past 2 decades have gotten rich thanks to a lucky pitch without a clear path to a viable product. At least 6 of those I know did so at the expense of developers that accepted equity over cash...and the developers wasted a ton of time and 2 even were briefly homeless as a result...and none of them live in California.
Hopefully wannabe senior leadership will try and take advantage of AI without taking advantage of developers, because most of us just want to write code and build something great.
SaaS dying is mostly nonsense.
Which is easier to vibecode - AI agent or Salesforce?
The fragmentation in the AI agent space will be markedly larger than at the base CRM layer.
Abd the AI agent is replaceable in under a minute but your data in Salesforce isn’t.
Reminds me of the story of when the Surgeon General (in the US) reported that smoking causes cancer.
People stopped smoking immediately, and cigarette sales tanked. The cigarette companies laughed (with all the phlegm in their throats and lungs) and sales came back 1-2 weeks later.
I suspect in a few months or a year companies with vibe-coded replacements for SaS products will find they need to go back: But, just like how many less people smoke today than in the past, the writing is clearly on the wall. At some point someone will figure out how to replace SaS with AI; it's just going to take a lot longer than many think.
> it's just going to take a lot longer than many think
I'd be very surprised if devs were fully replaced by AI in less than 10 years.
And where are they now (cigarette companies)?
https://straitsresearch.com/report/e-cigarette-market
Altria has 72.2% gross margins. 45% net
Well, that was 1964.
Smoking survived.
At least 2 have 100B market caps.
I was reading through the article and waiting for the key info to drop, but nope. It never did. It seemed like marketing fluff. If anything, vibe coding may eliminate some of the B2C SaaSes, but not B2B. If you think an enterprise is going to vibe code a B2B offering that they pay millions for, you're out of your mind.
Here's my general mantra regarding AI: NEVER take suggestions about AI from people who have a vested interest in it. CEOs of companies that train and offer LLMs, Authors of Books about LLMs and AI in general, etc.
This may come off as an unpopular opinion, but this is how I felt after listening to Steve Yegge recently. He has a new book about Vibe coding and he goes on in the interview/podcast to say that the best programmers he knows in the world (the ones better than him and maybe even the top world class programmers), would be equivalent to those of interns in an year, if they don't start vibe coding or use AI. I respect the guy, but damn, this is just peak delusion. He didn't even say it as a hyperbole, he meant it.
According to popular CEOs of companies training LLMs, 2024 was supposed to be the year that would eliminate the need for Junior and mid-level engineers. 2025 happened. Now, we are in 2026.
So yeah, I'm never taking advice about AI from these people ever again.
> Here's my general mantra regarding AI: NEVER take suggestions about AI from people who have a vested interest in it
I get where you're coming from, but let's say you're talking to a HVAC installer, and he recommends you a system to get - I'm sure there's financial self-interest on his part, but I do like to think that he knows quite a bit about what he does, and believes what he's selling is genuinely good stuff (and has reason to), even if he oversells it a bit.
The analogy can work if you're not looking for an HVAC at all and the HVAC guy is instead approaching you, unprompted, to explain that you need to buy this new system. Because if you don't, your business will become uncompetitive and fail.
True. That's the case with almost any commodity in life. That's why I was specific about AI.
The difference is, in other sectors, there's no fear-mongering. If you don't use their HVAC, it's fine. Your job isn't getting replaced. The air you breathe in your home isn't going to be fully polluted. You have other options.
With AI though, there's no middle ground. You either use their tool and become extremely successful (so much that you don't know what to do with that much success) or you're out of a job and become obsolete in like the next 3 seconds.
Are there real documented cases of a company replacing their SaS with a vibe-coded version?
Like I can see how a very small company could replace a portion of an overkill and underutilized SaS platform.
I don’t see how a larger more complex business could replace their SAP or ADP with a vibecoded version.
These stories are all very similar in where the author knows some CEO of an obscure company who told them they had an engineer reverse engineer and vibecode some obscure SaS solution and saved them $50K.
Customers don't want to pay for rolodex + 8000-data-point-visulization + integrated-automatic-contact-form-filling + assisted-real-time-closest-contact-scanning + 15 new-features-months + Annual-UI-rewrites + All-that-amazing-businesses-functionality-for-only-$199/usr/mo.
They want AI to vibecode a rolodex. And just use that. For 30 years if needed. 2000 LOC and a one time $20 cost.
The cancer of SaaS cannot die fast enough.
Any competent programmer can write 2000 lines in a few days or less. If the SaaS could be replaced by that no AI would be needed.
With a new agentic-lashup tearing across the internet every week, pointing the way to "gradient descent" software development, any purchasing manager worth their salt is going to ask some serious questions about their enormous SaaS bill before committing to another expensive long term contract. It follows that valuations must decline. Even if only because risks to moats have increased, but also because it makes sense to negotiate hard on pricing when there's fear in your counterparty.
Preposterous. Have you ever worked for a company as a programmer or for that matter as a manager? They don't just replace products ad hoc. There's an enormous amount of due diligence that goes into any new software product - making sure it fits the company, that it's secure, that it works properly... I recently worked at a small startup who spent on sales force @ $100,000 a year. I said you know what we could do this ourselves and you know what they said as every company I've ever said that to says? "We don't want to support it. We need to cover our ass. Everybody knows how to use this"
I'm not one to sell generative AI short at this point, but this seems at odds with the longer-term trend towards more centralized, off-the-shelf solutions, like Shopify, Salesforce, Squarespace, over the "bespoke," in-house alternatives that many companies developed in previous decades. Perhaps AI is making the TCO of "rolling your own" so low that it makes sense to go back to that over SaaS?
AI, as do most things, help the big players get bigger. If someone is automating small parts of the b2b layer they get dropped, but it’s harder to drop an automation that companies are used to. I don’t see how AI is changing that, companies spent a lot of time and money to set up the automation because it’s needed and because they can write a potential replacement cheaply doesn’t mean they are going to rip away something that works and is reliable.
Are B2B sales actually impacted or is the stock market just randomly predicting AI will impact B2B and selling off?
Since when does stock price / valuation have to match actual business realities?
I would assume one major thing here is that many orgs only need a small subset of functionality from what most products provide. Many times, that small subset of functionality is only "good enough" in and of itself, but the org is paying the premium for the entire suite of whatever it is. This makes realizing that an LLM can get them to MVP and beyond much easier.
Charging hundreds of thousands if not millions per year for very basic functionality is what is "killing" b2b SaaS.
There is also the benefit of being able to use a single database (and hence schema) across multiple "apps". In many cases the complexity arises from the fact that all these apps have their own databases.
Well it definitely killed mine so I can't say this is not true
> Well it definitely killed mine so I can't say this is not true
I feel like there's an interesting story in there.
I'm sorry to hear that ... if not too painful, would you mind sharing more (so others can learn).
Oh no...
I predict the fallout from this is companies being nickel & dimed to death by a million smaller subscriptions (rather than just cutting a huge check to Workday, ServiceNow, SAP, Oracle or similar). There is such a glut of AI ISV startups that are filling highly specific niches, and some are quite good, but they're all usually in the $10-50/mo/user. Gets to be big numbers pretty quickly in a large enterprise.
There will be an exiting explosion of internal apps and tools and then it will die down as companies get back to business. All those tools and apps will turn into legacy apps and they will eventually try to migrate data off them into some saas platform because its not core to business.
Saas will have to drop prices to be competitive so fat margins will go away which will probably affect margins for AWS etc.
Sure, vibe coding has impacted user's expectations. They know you can ship a new update easier and faster than before - and you actually can.
But, not sure which successful SaaS companies just stopped shipping any updates to the product, never talked to their customers and never added any new features to win over major new accounts - and still managed to survive and thrive?
And the author actually confirms this:
> AI isn’t killing B2B SaaS. It’s killing B2B SaaS that refuses to evolve.
> Sure, vibe coding has impacted user's expectations. They know you can ship a new update easier and faster than before - and you actually can.
Can you though? With major bugs? We've been getting more and more crashes, downtime, issues etc lately and a lot of it has had to do with vibe coding.
The whole point of these B2B SaaS is meant to be quality.
i.e. it's set users' expectations but in the wrong way.
> They know you can ship a new update easier and faster than before - and you actually can.
And all of those updates are just AI features.
>They know you can ship a new update easier and faster than before - and you actually can.
How's that going for Microsoft?
https://www.windowscentral.com/microsoft/windows-11/2025-has...
It's this generation's "build vs buy." I imagine it will play out the same way, like a revolving door. Customers churn because they can "build it themselves," then a year later when they're sick of maintaining a mess of code for some internal system instead of delivering value to their own customers, they come back. A blip.
Or maybe there is look at true value proposition of the B2B SaaS products. It is not big spend per user, but it does add up eventually. And then savings start to look like big numbers. Big enough for some manager to justify action. This might lead cutting some seats just on cost basis.
I can see three forms of competition here:
- A company vibe codes their own app to replace a SaaS. Great when they only wanted a small chunk of the functionality. - Startups benefitting from AI coding are copying mature SaaS companies and competing on price. - Mature SaaS companies are branching out into each others domains. Notion is doing email. Canva is doing an office suite.
here's the secret saas can vibe code features too on top of their paid well developed and secured api. they can get off their ass and vibe code a mcp wrapper, so user can use the ai tooling they pay for to interact with their saas. and they'd be called visionary hero of the agentic revolution.
but they don't want to. and they will be replaced, as it's good and well.
Some founders are realizing this, we're helping a lot of B2B SaaS achieve exactly this with our whitelabelled solution.
well then, happy first billion, where I send a resume?
Maybe the type of SaaS that's akin to stock media (photos, video, music). Just hard enough to do from scratch, but not important enough that it needs to be exceptional in it's field. I've made some money off software like that, and it was nice, but I always knew it couldn't last. Better developers took most of it from me years ago.
Anybody who says this doesn't understand build vs buy, and why companies buy in the first place, or they'll selling AI.
Bingo, the OP is selling AI, and wrote multiple blog posts to promote his tool, this is just fear induced sales pitch.
there is no saas downturn caused by AI. wall street is just starting to say hang on a minute, why is this SaaS stock trading at a price to earnings ratio of 300?
then the sell-off is attributed to AI because it is far easier to say to shareholders hey we know our company lost half its value but thats actually a good thing because we need to pivot to AI and we're going to spend all our free cash flow on AI software and our stock should totally be trading at 300x earnings again in a few weeks. if you can last another few months as CEO and the fed cuts rates you'll be able to ride it out
of course, the tide is going out on a few dogs. I don't think adobe will become dominant again
you see the same trend with mass-layoffs being blamed on AI. easy way to sell bad news to the shareholders
in 2026, AI and JE are the two reasons for absolutely everything
Something notable for SaaS which this article doesn't mention is that in some cases the reason to buy rather than make yourself is due to needing to handle a bunch of different regulations which LLMs don't threaten (barring businesses which would rather have lawsuits than pay for a SaaS).
We will have to go through the stage of disillusionment of what AI is and understanding of what it is not. There is too much FOMO and too many stories driven by the heavy-AI-invested parties today to see thing clearly.
The real story isn't that some enterprise mega corp is going to vibe-code their own Workday.
The real story is that SOME startups made up of one person (or small number of engineers) will do it, and create pricing pressures against Workday, or DocuSign, or other B2B SaaS.
But if you are a mega corp, spending 0.1% of your CapEx on SaaS subscriptions, do you really want to switch to products made by some no name one person startup? You might save 0.05% of your CapEx, but if something didn’t workout, your whole business will be screwed.
To your point, I think there are 3 main categories:
1. Too big to fail (SalesForce, ServiceNow, ServiceTitan, Shopify). They’re targeting megacorps’ core business operations. Switching costs are too high. They will survive,
2. B2B non-core (PagerDuty, Vanta, Monday, Atlassian). They’re going to have stiff competition and most here will fail or merge/consolidate. They have the most to lose because they’re non-core to a business’s success and pricing pressures will cause many of them will be easily vibecoded with enough time. The large TAM here will attract hundreds of competitors each.
3. Consumer SaaS (Notion, Canva, Grammarly, Dribbble). They're good as dead and can be buried.
It is obvious.
A middle 100-500 heads firm don't need enterprise level SaaS, a vibe coded website will suit them better.
Fundamentally, those workflow/orchestration SaaS needs to answer the question why people should pay you premium while only getting 80% where they want to be.
For the most part, you can replicate any B2B SaaS product in a spreadsheet. The same reasons why spreadsheets didn't kill B2B SaaS apply to "in house vibe coded SaaS replacements." The original in house apps are (and continue to be) spreadsheets.
> build once, sell the same thing again ad infinitum, and don’t suffer any marginal costs on more sales.
Unless you consider customer acquisition cost. Not considering cost of sales is one of the big mistakes software developer entrepreneurs make.
It kills low hanging fruit SaaS that was always just some minor piece of functionality wrapped in a subscription model usually based on 99% open source and 1% actual contribution.
I think people here need to accept that software is becoming electricity, you get charged when you use it and by how much. You don't pay for a box shaped electricity or purple color electricity, it is just electricity.
Sounds like a lot of products will be in trouble if AI becomes more advanced with producing secure code with maintenance processes to keep evaluating code on an ongoing basis
How corporate incentives are aligned will also define the trajectory. The person who is going to take the call whether to go for vibe coded tool or external vendor has to have enough incentive to put this neck on the line.
Imagine this, you are VP of finance. You know you can get a nice tax calculator built quickly using vibe coding, but will you put your neck on the line and say, let's replace the existing vendor and use my vibe coded tool. You might fired if you send a wrong tax report to your auditor.
Let's take another example - VP of sales wants performance report and visuazliation of the sales team. He has 200 BDRs. The daily sales standup depends on this report, team gets yelled at or praised basis this. Now, will this VP be willing to put his neck on the line and say - let's use my vibe coded report and discard the existing SaaS. Even if such a report feels trivial, it is vital for functioning the sales team and hence, it needs to be reliable.
I think vibe coding will work at prototype level - the trust gap is very huge right now to even consider it for internal tools.
And, until vibe coded tools are stress tested enough this trust gap will not be fulfilled. Think about this, why some of the biggest companies in the world still run on softwares built in 2000s.
The reason for divergence is actually much simpler. NASDAQ 100 includes data center builders, Morgan Stanley software index doesn't. Stock market is going down across the board if you exclude data center construction.
Remember when businesses ran on cobbled together access databases and vb? It was easier than building something ny prompting an llm.I made a good living just rewriting those things for them when they fell apart.
Can't wait for the "AI Is supercharging SAAS" article when the stocks go up, and another "SAAS IS DEAD" when they go down again
This feels like it started out being written by a human
but around the point "2. Security, authentication, and robustness" the LLM took over and finished it...
This was the same hype around NoCode a few years ago. Thing is people for software as a service and internal teams dont want to do the service part. A common thread among all of the vibe coded stuff I hear is that nobody talks about the next requirement which was handled. Frankly UIs are learnable and after a few tries everybody gets used to it. The extra polish is totally not worth it. Eventually every SaaS on the non-critical business path gets ripped out with one of the platform players such as SAP, SalesForce etc which are infinitely customizable. Salesforce even has its own language in which customers write their apps.
The system of record bit is quite correct and imo the only durable advantage. But that doesnt explain why Klaviyo shoyld worry. In fact marketing workflows are some of the most important systems of record directly tied to business earnings.
The assumtpion that in SaaS you build once and everybody uses it is false. There are always customizations or features that need to be built for big contracts. Its the small players that are okay with out of the box SaaS since they lack the budget to pay for a customization. Ironically, these teams will get hurt the most should they choose to go down the vibe code everything path. Here infra capex is far lower than payroll and owning too much software (vibe coded or not) will simply steal time from business development.
Honestly, the market just panicked and caused a temporary correction thats all. Then we all got busy and wrote a bunch of thought leadership crap on top.
In reality AI is the best thing to happen to SaaS companies since it reduces RnD time and expenses, allows even small players to bid for large contracts without overloading existing dev capacity and also makes it possible to put devs in front of customers since now there's no need to sit down and code. AI is also the best thing to happen to engineering since it now allows for product management to be folded into the existing craft.
Be a System of Record, not just a Wrapper™ is excellent advice.
Thank you! I hope you enjoyed the little ™ at the end as well.
Previously on HN:
https://news.ycombinator.com/item?id=46268452
AI isn't killing B2B SaaS. It's killing the service economy. Perhaps, the correct term, technically, is just shrinking it to very very small fraction.
I wouldn't see it positive that we just need to rethink and offer platforms. As this platforms already exist Azure, AWS, GCP
Maybe things will finally go full circle and people/companies will restart self-hosting their infrastructure instead of farming out everything.
SaaS’s competitive advantage is not tech but distribution.
Just because the stock market dips, and a bunch of media outlets ascribe this to AI doesn’t mean that AI ‘caused’ the dip.
If people really think non-tech businesses will start running their own software organizations and maintain all these tools themselves is an idiotic assumption at best.
The “building” of software has never been and will never be the bottleneck.
The trick is to build stuff that is hard to vibe code
Yes, that is not the random SaaS product, possibly not any SaaS software at all.
Its not that hard.
I am building a desktop app that deploys and manages high availability postgres instances on VPS right now..
The AI is constantly failing at helping me because its about generating and managing remote configuration that is very fragile if something is missing.
Build infra heavy saas. Somehow can get through. Llm will get better only, and people start copying every sucessfull platforms.
I don't think AI is killing B2B SaaS so much as companies are finally reckoning with the immense costs of SaaS in a markably different environment than when SaaS exploded in popularity, and AI offers an off-ramp to some. Let's break it down camp-by-camp to show you what I mean:
1) The must-haves. These are your email and communication systems, the things you absolutely have to have up and available at all times to do business. While previously self-hosted (Exchange/Sendmail, IRC/Skype/Jabber, CallManager/UCS), the immense costs and complexities of managing systems ultimately built on archaic, monolithic, and otherwise difficult-to-scale technologies meant that SaaS made sense from a cost and a technical perspective. Let's face it, the fact nobody really hosts their own e-mail anymore in favor of Proton/Microsoft/Google/et al shows that self-hosting is the exception here, not the norm - and they're not going anywhere regardless of how bad the economy gets. These are the "housing stock" of business, and there's plenty of cheap stock always available to setup shop in without the need for technical talent.
2) The juggernauts. The, "we can do this ourselves, but the pain will be so immense that we really don't want to". This is the area where early SaaS solutions cornered and exploded in growth (O365, ServiceNow, Google Workspaces), because managing these things yourself - while feasible, even preferable - was just too cheap to pass up having someone else wrangle on your behalf with a reasonable SLA, freeing up your tech talent for all the other stuff. The problem is that once-focused products have become huge behemoths of complex features that most customers neither need nor use on a regular basis, at least after the initial pricey integration. Add in the ease of maintainability and scalability brought by containers or microservices, along with the availability and reliability of public cloud infrastructure, and suddenly there's more businesses re-evaluating their relationships with these products in the face of ever-rising prices. With AI tooling making data exfiltration and integration easier than ever from these sorts of products, I expect businesses to start consolidating into a single source of truth instead of using dozens of specific product suites - but not toppling any outright.
3) The nice-to-haves. The Figmas, the HubSpots, the myriad of niche-function-high-cost SaaS companies out there making up the bulk of the market. Those whose products lack self-hosted alternatives risk having vibe-coded alternatives be "good enough" for an Enterprise looking to slash costs without regard to long-term support or quality; those who compete with self-hosted alternatives are almost certainly cooked, to varying degrees. If AI tooling can crank out content similar in quality to Figma and the company has tech talent to refine it for long-term use, why bother paying for Figma? If AI tooling can crank out a CRUD UI for users that just executes standard REST API calls behind the scenes, then why bother paying for fancy frontends? While it's technically interesting and novel at how these startups solved issues around scaling, or databases, or tenancy, the reality is that a lot of these niche products or services could be handled in-house with a container manager, a Postgres instance, and a mid-level IT person to poke it when things go pear-shaped. The higher per-seat prices of a lot of these services make them ripe for replacement in businesses comfortable with leveraging AI for building solutions, and I expect that number to grow as the tools become more widely available and IT-friendly in terms of security.
Ultimately, the core promise of SaaS to business customers was all the functionality with none of the costs of self-hosting support. Nowadays, many of them have evolved into solutions that are more expensive than self-hosted options, and businesses that have shifted IT into public clouds or container-based systems have realized they can do the same thing for less themselves, at the cost of some UI/UX niceties in the process. Now that we (IT) can crank out integrations with local LLMs with little to no cost, we're finally able to merge datasets into singular pools or services - and I'm not talking about Snowflake or its "big data" ilk so much as just finally getting everything into Salesforce or ServiceNow without having to bring in consultants.
The must-haves and many of the juggernauts will remain - for now. It's the niche players that need to watch their moats.
Anyone can learn to cook, and the barrier to entry is low. There's no shortage of restaurants though.
There's no shortage of restaurants that go out of business. Even big name brands have issues with location turnover.
That's a great point. The restaurant business is one of the worst you could possibly go into, with razor thin margins and astronomical failure rates. If those are the future economics of SaaS (accustomed to 70-80% gross margins), than you ought to cash out now while you still can.
Slight tangent: what’s the deal with the weird ligatures on that page in the headings?
I'm the biggest typography nerd and love ligatures. Can't tell if the majority likes them or dislikes them lol
People have been debating what "real" development is for as long as there have been computers. "That's not real software development, you're not even controlling which registers you use!" "How can you say that's real code when you aren't even managing your own memory!" "That doesn't look like C, must not be code!" and on and on.
At the level of the old adage about whether the horse-drawn buggy-makers are in the buggy business or the transportation business, it's all telling the computer what to do in the context of providing a customized tool that you or others might use. So in this context, a customized Excel spreadsheet counts. And so does a vibe-coded app.
And of course, wringing our hands about what it looks like now totally ignores the fact that it's not going to be like this for more than a year or two at most.
How long until a user can reasonably say to Claude or similar, "I need Bob here to track production at my factory. Lay out a set of tools to do that, and make sure they're layered with help and tutorials so Bob can learn on the job because he doesn't know anything. Don't let him make any mistakes."
That's probably not coming next year, but certainly it's not ten years away.
The problem is never building, it's maintenance. SaaS isn't going anywhere.
Maybe the new SaaS is to build vibe coding (aka conversation) into whatever you’re offering.
thats what theyre all doing. its gimmick and wont really fix the underlying issue.
Having worked in enterprise B2B SaaS for a long time, almost every feature I built could have been a simple spreadsheet or some emails. So I'm highly skeptical AI is going to change anything.
Enterprise sales basically works like this: A non-technical sales team aggressively promises everything to win a deal to a non-technical procurement or exec team. When the deal is won, the SaaS sales team tells engineers "go build this" regardless of how stupid it is. And the customer tells their employees "you now have to use this SaaS" regardless of whether it makes sense.
AI is killing this. Millennials are killing that. Gen Z certainly must be killing another thing. And Yarn? It was killing NPM. This nonsense type of title really was made for a world where nuance doesn’t exist.
B2B SaaS is projected to grow over the next decade, not shrink. Just use the LLMs and you’ll be reminded of the value provided by the company selling non-core but important tools for your business.
no. High interest rates and a cautionary view of future economic growth are killing B2B SaaS. Money is no longer free, and so there is a bigger push for cost-cutting rather than growing your buisness with free money.
A link shortener is such an easy thing to code, it's essentially one database table with a redirect. To add to that, there are many open source libraries to implement link shortening, including analytics and stuff. Even then Bitly and Rebrandly have customers (from their website) like Toyota, Cisco, Oracle, Monday.com, New York Times, etc.
Are these companies unable to build a link shortener? It's also so easy to migrate off shortener service. If they can and still choose to use these shortening services, there must be other reason. And that reason is that they simply don't want to. This has nothing to do with AI.
I run a software company and one of the reasons customers say they want to migrate from their homegrown spreadsheet is because the guy who built it left. A freaking spreadsheet!
Such blog posts and probably many comments here are the perfect answer to "Tell me you don't run a real business without telling me you don't run a real business"
Regarding your last comment, majority of the people here are costal with FAT paychecks slinging code for VCs. It’s a totally different universe than running a Saas. That said, still a valuable forum.
Until Claude Code comes with indemnity insurance for HIPAA / GDPR / etc… B2B SaaS is here to stay. You want me to convince my auditor that the vibe-coded in house software handles PII correctly?
Making the audit someone else’s problem is 90% of the ‘buy’ value in ‘build vs buy’
Spot on. You could argue that most companies buying B2B SaaS could almost always build a clone internally but they need someone to assume SLA and liability.
Jellyfish is a Jolly Fish, also invertebrate
I know this is petty but I stopped reading when I saw the “c-t” ligature in the article headings. Obnoxious and pretentious.
I'm the biggest typography nerd and I'll fight to death in the defense of ligatures! Bring them back
I love ligatures and typography too but not these.
I wouldn't say that AI is killing B2B SaaS, let's analyze the situation: AI is shifting the boundaries of the Build VS Buy decision.
If software is cheaper... building it seems a much better option but... ...the fact that software is cheaper means that SaaS companies will be able to be more competitive with price, offering more features, integrating them much better with the newest agentic tools and frameworks that are getting released.
Sure, the market will adjust, some will be pushed away, others will find businesses that weren't possible before... but we're far away (I hope) from the AGI revolution, don't forget to see this crysis as an opportunity too.
That’s cool. I’m building a mob programmer.
Vibe coded SaaS is a SaaTA. Throw away.
the procurement bypass was the best part. now watching ai devs ship faster than our salesforce admin could configure flows
I've worked in SaaS for most of my career, only recently working at a big corp who is largely the buyer and user of SaaS tools to meet their objectives. From the perspective of the corp business buyer, they want something that works for their needs and they want to buy something instead of build it because the support costs are gnarly. They already have engineers dedicated to the tools they've purchased. Much better to put the risk on someone else they can yell at. And the permissions and access to these tools, reports, data, is usually its own special problem to manage. Building a lot of one-off tools is going to just give IT a huge headache and they will push the org to buy before vibe coding a solution.
if you're a software company and all your clients are in tech...you're gonna have a bad time. godspeed.
Look, software is not going away even though everybody thinks they're a developer now. Do you think companies are going to replace Microsoft Windows, of which there are a billion installs, or Salesforce, Crowdstrike, Quickbooks etc, with some Vibe coded AI slop an intern "coded"?
When enterprises/businesses in general upgrade any software in the company it takes years sometimes... There is also Vendor lock in, you can't just swap things with vibe-coded slop, and trust me your manager will never want to do that either because his butt will be on the line.
I disliked how SaaS CEO's were decrying the death of engineers. Their coordinated layoffs over the past years or so was excruciating to watch and experience. Their language was aggressive and inflammatory.
Although the article may also be hyperbolic, I'm not going to comment on reasons why it might be. Instead, I will agree, and think SaaS companies stock performance this year will be proof. Sure, it might not be the collapse that AI doomers are hoping for, but all the FUD they spread over the past few months to years will signal that they're not insulated from it. They made their cake, now they have to eat it too.
At the end of the day you do not need to replace your B2B SaaS with AI.
You need your B2B SaaS to think you can use AI to replace it though, so said SaaS will keep it's prices reasonable. Otherwise they have you by the balls and will charge you much as humanly possible.
been thinking about this too. the middleware layer is getting decimated first
If that would be true, expect in the next decade a frantic search for seclusive grey beards, those who haven't given up their rituals and ancient languages.
If your workforce is vibing all day, they will have no capacity for maintenance, because it isn't their code. So the maintenance that happens will be slop and more spaghetti. I am not saying cases like that never existed before, but such companies will face a moment of truth sooner or later.
The survival advice here — 'be a system of record,' 'make switching painful with compliance' — is exactly the playbook that makes customers want to leave in the first place. The deeper question is whether vibe-coded replacements will rot fast enough to send people running back, or whether AI gets good enough to maintain them too.
No it isn’t. Writing the code was never the issue with making software, it was designing it.
You can shit out an app with AI, just like you could with Indian workers. But that doesn’t mean it will work properly or that you’ll be able to maintain it.
And most importantly, it only works for code they could steal from GitHub. It has no idea how to replicate sensitive systems which aren’t publically documented, and those are some of the most valuable contracts.
> how they’re potentially losing an $X00,000 account just because the customer can’t use a specific failure reporting workflow in the SaaS. They’re now working with me to build what the customer needs and retain them.
You mean the poor SaaS company actually now has to implement features needed by customers??
Jesus wept.
at last, TrueAnon has arrived at hackernews
Just a single data point, and I‘m not pretending it’s a conclusive one (yet), but I think there is a middle ground between buying a SaaS and vibe coding a replacement: replacing a SaaS with your own solution, using AI coding agents — while actually knowing what you do because creating robust software in-house is already a core competency. No vibe coding needed.
At my company, we build software every day because our business is running a job board.
We always had kind of an impedance mismatch when it comes to creating content pages (think landing pages for marketing).
Yes, we can do this ourselves, but then software engineering resources are in conflict between shipping the next feature and creating landing pages.
So we introduced Webflow. Now marketing could create their content themselves. Did it match our corporate identity? Hopefully. Was it technically sound? Sometimes. Was it fun for software engineers to fix things in Webflow. No.
It kind of worked, but wasn’t exactly ideal.
Meanwhile, software engineering became more and more productive with the advent of AI coding agents, Cursor in our case.
So we tried another approach: giving our content creators Cursor.
But that was brittle, too: Cursor presents an overwhelming complex UI for someone who never used an IDE before, it could do a thousand things while only three are actually needed in this use case, you have to explain git on top of Cursor. It kind of worked, but only kind of.
So we sat down and built https://dx-tooling.org/sitebuilder/
It’s like a hyper-focused „Cursor light“ in the browser, so our content creators can just „chat away“ and create their content pages, with all the guardrails baked into the product. Getting tracking pixel integration etc. right just works. Matching our style guide perfectly just works.
And as a bonus, there is a whole git based storage and versioning workflow underneath, so software engineering can support and test and deploy the results with their preferred tools and methods, but none of this complexity leaks through to the content creators.
We built this ourselves in days, not months, thanks to Cursor, but it’s not vibe-coded — it’s a rock solid application that we will support and improve long-term without headaches: https://github.com/dx-tooling/sitebuilder-webapp
Time will tell.
No.. just.. no. This will be a thing for like 1, maybe 2 years, then people will realise it doesn’t make sense to spend $50K of time per annum to replace a $500/month subscription for a better product.
One point is now you dont have to pay money to 3rd world countries like mine :(, you can manage with fewer resources inhouse. AI will take care of slop work.
for profit software is pretty gross tho. it can be made indefinitely complex, and people can make and sell tools to patchwork the complexity together, but then u need more tools to patchwork all the tools together ...
and u end up in aggregators aggregated aggregators type situation where optimal solutions never arise because we don't actually cooperate enough to produce them
ai is fitting into the notion that this is all bullshit ... even if not in the correct way
Article is AI slop based on a mostly made up premise.
goooood.
“Software is eating the world” and “AI is eating software”
Honestly, I'm surprised by how people are pushing back against this idea. I feel like vibe coding is just in its earliest moments of actual viability, and my mind is totally blown by it, and it strikes me that it's exactly what I've always wished software could be. Plastic, flexible, personalized, effective, responsive, organic.
As an anecdote, I've been vibe coding an accounting system that perfectly matches with my own expectations of accounting software, i.e., it's intimately connected to CSVs, imports and exports from CSVs, but acts as a kind of enrichment and reporting and file association layer. If there was anything like this, any kind of SaaS that I could have and download as software and run on my own computer offline and be able to inspect and trust and version control so they wouldn't add or remove some kind of feature that I wanted down the line, then I would have gone with that.
But now I have essentially my absolute ideal solution, written with a Python backend and Vue and JavaScript frontend, and it's radically improved my ability to maintain accounting for our business account.
And I think there's something really important to point out here, which is that the feeling of lock-in is very seriously reduced when you are Vibe coding your own software, because if you don't like the way that it works or you realize that there's something missing, you can add it pretty painlessly. Like, that's always been a huge challenge with choosing vendors for a SaaS platform, is you think, oh no, well, what if their conceptual model for what I'm trying to do doesn't quite map onto our own internal systems or understanding of what's being done? Well, when you have your own Vibe-coded SaaS, you can just add that information. So there's something incredibly organic about it. I used to work at a startup in Redmond where we built this large internal system to manage a scientific process with lots of machinery and data, and it was incredibly empowering and actually became one of the core values of the business that was able to be licensed to other businesses in the same industry. And it seems like we're just improving that capacity from here.
Now obviously, if Vibe Coding magically were to go away or became much more expensive, then I'd have this legacy piece of software, which I couldn't improve, and that would be a dead end. But my expectation is that the functionality that we have today will only improve. And in several years, the scope of changes that I'll be able to make, the level of professionalism, modularization, maintainability, code quality, will only improve. And so this has me thinking in general that software is kind of undergoing a step change where we're moving into the so-called age of intent beyond the age of the interface. And that's tremendously exciting to me, and I just couldn't be more stoked about it.
[dead]
This immediately lost credibility for me with this quote:
> And vibe coding is fun. Even Bret Taylor, OpenAI’s chair, acknowledges it’s become a legitimate development approach.
Color me shocked! Bret, who directly profits by how his product is perceived, thinks it's legitimate???? /s
And if I understand correctly the author is running a business that helps SaaS companies overcome the risk of using their own vibe-coded solutions.
Good point -- removed for being biased and partial. Thanks for the feedback!
> Good point -- removed for being biased and impartial. Thanks for the feedback!
??? Do you mean biased or do you mean impartial?
"biased" and "impartial" are antonyms. Pick one or the other.
1 reply →
[dead]
[dead]
[dead]