Ask HN: Why LLM providers sell access instead of consulting services?
4 days ago
I don't understand the business model of AI companies like OpenAI and Anthropic: if the claims of nearly autonomous agentic product development are true, why would they sell access to their LLMs, in some cases at a loss, instead of providing consulting services at a fraction of the cost when compared to traditional human businesses?
In other words, why would they sell AI as a sort of commodity instead of using it to provide an end product like IT consulting which is extremely profitable?
Because APIs scale but consulting doesn’t :)
Even with good models, delivering outcomes is still messy. Bad data, unclear requirements, integrations, and blame when things break. That’s just consulting pain.
Selling access avoids delivery risk and headcount bloat. Otherwise they would become a service oriented company with AI. APIs give global distribution, cleaner margins, and optionality. They can always move up the stack later
I work in consulting cloud + app dev. Every single project I’ve had for the last year and a half has been Gen AI related in some form. Most of them could have been done before Gen AI using traditional ML. It’s just much easier now.
That being said, consulting is a low margin race to the bottom business if your consultants email address isn’t @amazon.com (been there done that) or @google.com.
Google’s sales staff and SAs and other go to market to chase large deals are still not on par with AWS ProServe or Microsoft’s. It takes a lot of work to build up a good consulting department and no matter what, you still end up - unless you are a cloud provider -going with the Americans are the face for the customer and low paid folks from India do all of the work.
Consulting for cloud companies - either internal or external (where they give credits to customers for third party partners consulting) isn’t about the money they make on consulting, it’s about ongoing spend on their cloud providers and they can also afford to lavish credits on companies. The cloud provider have real profitable businesses
With consulting you make some $X * N consultants on staff
With products you hope to make $X * N units sold. Staffing to build product + make those N sales takes something like log(N) sold units.
At least that's the dream. Consulting requires little upfront capital investment. Products often require a constant, but large, upfront capital investment
Because selling sophisticated shovels in the gold rush is much more profitable than finding gold and trying to make and then sell jewelry
Traditionally yes. But not when your business model (if you can call it that) is selling shovels at a loss to wholesalers (cloud providers) who are making all of the money selling it to customers. AWS, GCP and Azure are all making a killing and the only reason they aren’t more profitable is because they can’t get enough chips.
Microsoft just said as much in their quarterly filings and unrelated to AI, Cook said they couldn’t meet all of the iPhone demand because of capacity issues at TSMC
yes, recursively selling tools for shovel makers can be even better :)
1 reply →
I saw a number earlier that around 45% of US businesses have a paid subscription to an AI service/tool. You just can't get to this kind of scale and revenue with consulting.
Consulting has weak margins compared to SaaS and scales poorly. Providing the interface for companies to spin up their own consultants (=Agents like Claude Code) is a superior business model in every dimension.
But those margins are for traditional businesses with human workers, if these claims of 100x productivity increase are real Anthropic should very easily be able to outcompete Accenture no?
Consulting - especially the more strategy type consulting - is often not about “we don’t know how to do something”, it’s more of “there is so much resistance to change organizationally that not even CxOs/directors can push it through”.
Besides selling consulting services involves a lot of relationship building and knowing the business vertical.
Yup, because LLM inference can be scaled by adding racks of hardware. Consulting can't be.
AI isn’t good enough to do consulting yet.
They get more data this way. I think their actual end goal is to supplant the software industry. I would not be surprised if there is an OpenAI OS in the next 5 years.
The problem is that they'll then be competing with the companies that are paying them now.
[dead]