The author observes that a call to GPT-5.6 Luna is only 4-5 orders of magnitude more expensive than grep, and then predicts that at current rates of progress, calling an LLM will soon be cheaper than a grep. I think this is a good time to invoke Stein's Law: "If something cannot go on forever, it will stop." These efficiency improvements won't continue forever. It's more likely that the per-call cost of high-quality, compiled software like grep will be a lower-bound that LLMs asymptotically approach, rather than a line that they blow past with perpetual exponential progress. (Barring a true breakthrough in something like quantum computing or room-temperature superconductors.)
You can burn anything* into an ASIC to make it cheaper per-call.
non-backreferencing grep is not very difficult to implement in an ASIC either. But it's probably not worth it because of how relatively rarely you use it and of the data transfer costs.
LLMs are great candidates for ASIC-burning because they're slow compared even to network speeds and run all the time. The issue is that you don't want to burn a specific model or architecture that then becomes obsolete.
So you've got two possible futures, and both guarantee large price drops: (a) LLMs keep getting better and better and better, so ability/$ keeps rising; or (b) LLMs plateau in ability, in which they will start getting ASIC'd.
depends on what you are grepping ... greapping a large file might be more expensive one day than generating n-th token with LLM that works fully in hardware
you could make hardware implementation of grep and store the file itself next to it in some ROM but that's not a very useful grep ... while hardware LLM is exactly as useful as software LLM only orders of magnitude faster
4-5 orders of magnitude is huge. Assuming an order of base 10, it's 10000x-100000x. So a call to grep may return in 1s on a typical PC. That means a GPT call takes equivalent energy of 10000-100000 PCs to do the same in 1s. That's a difference that can't be equalized with scaling. It would require a revolutionary breakthrough.
I also don't understand where the idea that frontier models are getting better efficiency comes from. The results are certainly improving, but that comes from feedback and multiplexing requests, which cost more.
LLM is spicy memoizing, so it can potentially be faster than a tool call. But people will spend a month tweaking and testing to ensure they have the level of determinism they need, which means it's more expensive, and that they should have used actual memoization in the first place.
> The author observes that a call to GPT-5.6 Luna is only 4-5 orders of magnitude more expensive than grep, and then predicts that at current rates of progress, calling an LLM will soon be cheaper than a grep.
At some future point where LLM hardware is cheaper than simply running grep, then grep equivalent would benefit from those selfsame hardware improvements and be cheaper to run as well, probably still by the same ratio.
But if the problem is literally grep (search this file you've never seen before), no index can pre-exist.
If you assume the file arrives ahead of time, can be indexed, and that this is worthwhile because we want to support multiple pattern matched retrievals, then sure it makes sense to consider indexed query schemes and upper/lower bounds. Each query could be faster as an inference if it doesn't have to re-scan the whole file.
But I don't think anybody, in good faith, can pretend that any LLM can digest a file faster than grep can. Particularly, if you admit the vector processing dedicated to doing the convolution kernel(s), you should also admit similar hardware could run a vectorized grep.
The LLM doesn't have an index of every single file I might want to grep, though. In practice it has an index of very few of them, and perhaps even none of them.
It might never beat out grep, but it could beat some more expensive to call tools, similar to how heuristics will often be faster than exact answers. Rust Analyzer can be slow at times, I could see an AI tool taking over a subset of its work.
It's probably better to optimize rust-analyzer first (and now, this became easier). I mean, see rust-glance: it's not feature complete but it points out to different tradeoffs in this space
A few years is hardly forever and the state of the world here indicates a lot of low hanging fruit still exists.
An LLM can certainly be cheaper than grep, because it’s an approximation, while a grep is deterministic and must examine every byte in what can be a relatively complex state machine for a regex based grep. There are other scales to consider like the scale of your local hardware vs the highly multitenant and high end hardware of the hyper scale inference providers.
There are already high volume models for coding inference where the reasoning time is crazy low and cheap per token where it can build reasonably simple software so blindingly fast it isn’t implausible the bottleneck is the latency in tools and networks. I find them hard to use at times because I don’t have time to think through the next turn by the time it’s done.
Regardless I wouldn’t be surprised to see a world where tokens are so cheap it’s not worth metering them but charging licensing feels with meter tiers at the far horizons to prevent abuse, charge outliers. Subscription models already set this stage well.
The other side to consider is bountiful capacity will also drive tokens to near zero price. The data center build out is barely underway and as it materializes, as hardware efficiencies improve, as techniques and model science and technology improves, harnesses, methodologies , etc improve, the economics flip from load shedding to trying to keep the data centers utilized. The economics lead to the world where tokens are not a unit of measurement for cost for anyone other than the inference providers to manage their utilization.
> An LLM can certainly be cheaper than grep, because it’s an approximation, while a grep is deterministic and must examine every byte
You mean a grep over terabytes of data vs a LLM with gigabytes of parameters?
If you have so much data, you can use an index to search. It's unlikely that LLMs are going to be cheaper than properly indexed search DBs (which is what we should be comparing them with)
There is no reason to assume that quantum computers would benefit LLMs in particular. Perhaps we could implement LLMs as analog circuits to save energy.
What if the tool is more advanced, like an optimizing compiler: `g++ -O3 -march=native -x c++ - <<EOF ... EOF`
If the compiler invocation is sufficiently slow, the llm could consider outputting a binary directly?
For all we know matrix multiplications are a faster way to generate optimized machine code than branchy sequential compiler code with tons of heuristics and passes.
From a computational standpoint this is obviously nonsense, but from an attentional one I'm not so sure. It may already be more attentionally expensive to use grep in some cases, such the moment you need to remember a non standard arg. And if this applies for performing a simple http operations, then it certainly applies going up the complexity chain.
room-temperature superconductors, sure, but I fail to see how quantum computing will disrupt – in the medium term (25 years or so) – classical computing in any meaningful way
Is running LLMs (or some other ML workload) on/with quantum computers expected to bring efficiency gains?
I think it's an interesting thought experiment: could an LLM call be a "cheaper" grep?
Especially for deterministic activities it just feels impossible to imagine general LLM tech handling the problem better, despite everything being said.
But hey, tech is filled with "smashing the generalist hammer works better than the specialized tooling". Would be odd though!
One thing that's easy to miss about performance is that it depends on framing.
An example I like to give: optimizing a data processing program's runtime by 5x is obviously 5x speedup for everyone. But if, for some reason[0], this means it crashes and restarts more often, it stops looking like this to end users. If every restart means it needs to start from scratch, and it restarts 10 times on average now where it didn't restart before, the 5x speedup suddenly looks like 2x slowdown to end user.
In this sense, LLMs are already much more efficient than most CLI tools, by a combination of:
- User not having to remember the exact invocation, or even the name of the CLI tools
- LLM being able to run the CLI tools and chain them on its own
- LLM being able to self-correct in case it got things wrong, or when actual output show that user's idea was right, but the instructions were wrong
Prompting "okay, list those processes sorted by runtime and match them against these output files" is both faster to type than the actual commands, it also end-to-end completes much faster than doing it by hand.
--
[0] - And I don't mean a bug. Say it's some batch processing run on a cluster with aggressive resource usage management; 5x speedup means it runs much hotter than before, which may put it on the top of "kill list" for when the cluster managing code needs to free up some resources.
I found the OP insightful and worth a read. Thank you for sharing it on HN.
The only aspect that is poorly analyzed by the OP is business model viability. All players are investing insane amounts of money in infrastructure with the expectation that their future profits will justify all that investment. The winner or winners in the AGI race, they believe, will find the proverbial "pot of gold at the end of the rainbow."
The OP glosses over questions of business model viability with a brief qualitative discussion and very little hard data. For example, to earn an annual return > 10% on every trillion dollars of capital sunk into infrastructure, the owners of that infrastructure must earn free cash flow (operating profit less investment) in excess of $100 billion per year in perpetuity. Is that feasible? Why? How?
I think the article's analysis is basically right in a vacuum. That is, I think it's clear that inference is a viable business model. But what isn't clear is whether it will be such a profitable business model for any given company that it will justify the investment that company has taken. I kind of think the winners might be a follow-on generation of companies that focus on this commodity inference business model instead of the invent-machine-god-first "business model" and thus are wiser about their level of investment and capital costs.
> I think it's clear that inference is a viable business model.
You may be right. I'm not so sure. Inference looks like a viable business model for those operators that have SOTA infrastructure in place, but the investment required to have it is enormous, and appears to be never-ending, because if an operator stops investing aggressively, its infrastructure quickly becomes non-competitive, and customers will quickly leave for alternatives. SOTA infrastructure is a moving target.
> I think it's clear that inference is a viable business model.
Only if you also have the model thats better than anyone else's.
As soon as models are free, or there are no newer models (assuming thats going to happen, and thats not a given) then the only thing you can compete on is price.
This means that the only thing you have to differentiate is either price, speed or ease of use. (or regulatory capture...)
We are at pets.com level of spend currently. Unless model development becomes cheaper, then we are going to run out of novel debt but not really debt mechanisms.
The problem source is that the "cost" of tokens are taken at face value from business that are losing money at record speeds. E.g. https://artificialanalysis.ai says "doing task A costed us $10 using OpenAI", and that is the "cost" the OP used as basis for "tokens are cheap". Meanwhile OpenAI is losing $19 for each $1 in revenue... So right now OpenAI should be charging around $200 to do task A just to break even, but that would mean their use base would collapse.
>The winner or winners in the AGI race, they believe, will find the proverbial "pot of gold at the end of the rainbow."
The hubris of this is really astounding too. There is no technology out there that some company develops and has not been reverse engineered and copied and manufactured at scale by competitors before long. You can't stop this from happening. People will leave the company or be poached and proliferate what they have built in the past. Every country that wanted a nuke has a nuke, after all.
The only way to keep the secret fire from leaking out would be to have AGI's first move be to lock the doors and prevent anyone from ever leaving company property again.
Some frontier labs are reporting positive "adjusted EBITDA" (earnings before interest, taxes, depreciation, and amortization, with extra adjustments to make the figure positive).
Free cash flow (operating profit less investment), actual cash coming in, is deeply in the red.
EBITDA can be a sensible measure of profitability when there isn't much need for additional investment. That doesn't seem to be the case with these operators. They need to invest aggressively to avoid losing customers to competitors. All of these operators have made multi-year commitments to invest more in infrastructure. In addition, they have guaranteed quite a bit of debt to fund it.
Maybe it all will work out fine (and I sure hope it does!), but I didn't see any hard data from the OP, or from you, supporting that view.
Definitely. The question is: Is it enough to recoup the enormous capital costs and justify the level of investment they've received. I think there's a decent chance that it will be. But maybe not. And the longer they keep focusing on training new models more so than on inference, the more uncertain I become that it's all going to work out.
Labs are playing money games with EBITDA, which is not uncommon, but also hides the extent to which they are in the red (deeply, deeply, in the red, and projected by them to get worse).
Think of it like spending $1 trillion to become the next Google. That hardware itself may never turn a profit. But if 10 years for now you're the software provider that owns the ecosystem around too cheap to meter tokens you've got a money printing machine.
Exactly. "Cost-to-distill" is a critical parameter. Right now usage of frontier models for all tasks is both subsidized and irrationally popular even at the subsidized price. Deepseek would solve most tasks faster and 10x cheaper. I agree with the author that just as Deloitte exists, frontier labs will exist. But not because their products are proprietary technical marvels or gods, but rather because of branding.
DS wouldn't be 10x cheaper than the subsidized subscription plans from openai/anthropic. Although it is of course much cheaper than the enterprier/API pricing- I think if you're on the subscription plans, you can't beat that on performance per price.
It's a deliberate reference/meme that is basically used to acknowledge the precedent of overly exuberant predictions of cost in an emerging technology but argue "however, this time it's true".
Of course, perilous territory for future irony depending on how your prediction plays out.
It's really quite unfortunate that the promise was not delivered, mostly for political reasons. I hope that a new wave of reactors and the dire need for clean energy restarts the nuclear race.
The economics are not there. Solar + batteries are much cheaper per watt today and still improving. Even better, the solar can come online instantly and expand while US nuclear takes twenty years to start generating any energy.
One excellent nerd-sniping side-effect of those political reasons is that the good folks of Austria built an entire nuclear power station and then never put it into service [0]. Open for guided tours on Fridays [1].
Nuclear was never going to give us too cheap to meter regardless of politics. Uranium just isn’t that cheap. Fuel costs are lower then coal or gas, but not so low that operators would just not bother to charge for it
Having your own solar can be "too cheap to meter", though it is definitely a feast and a famine kind of a thing.
We had batteries full and were spending electricity on all kinds of luxury things like whole-day internet and desalination for several weeks, and now that it has rained for over a week we're starting to turn non-critical systems off to keep the lights on.
On the other hand, this did work out in other areas. I pay a flat monthly rate for all-I-care-to-eat internet access, for example. My email provider has limits on storage space but I don’t get charged per email sent or received. It’s not a crazy concept on its face, nuclear power just didn’t work out as well as was hoped.
"Between digital economics and the wholesale embrace of King's Gillette's experiment in price shifting, we are entering an era when free will be seen as the norm, not an anomaly. How big a deal is that? Well, consider this analogy: In 1954, at the dawn of nuclear power, Lewis Strauss, head of the Atomic Energy Commission, promised that we were entering an age when electricity would be "too cheap to meter." Needless to say, that didn't happen, mostly because the risks of nuclear energy hugely increased its costs. But what if he'd been right? What if electricity had in fact become virtually free? The answer is that everything electricity touched—which is to say just about everything—would have been transformed. Rather than balance electricity against other energy sources, we'd use electricity for as many things as we could—we'd waste it, in fact, because it would be too cheap to worry about."
... What Mead understood is that a psychological switch should flip as things head toward zero. Even though they may never become entirely free, as the price drops there is great advantage to be had in treating them as if they were free. Not too cheap to meter, as Atomic Energy Commission chief Lewis Strauss said in a different context, but too cheap to matter. Indeed, the history of technological innovation has been marked by people spotting such price and performance trends and getting ahead of them."
My issue is that datacenter energy costs are being prioritized for commerce over residential use, so the average consumer is paying more for electricity, because a datacenter needs more electricity and they're getting tax breaks. Assuming this all improves efficiency for new products like automated robotics, there is a debatable benefit. Jevon's Paradox has no ceiling, except the environment, and people's 401ks.
I just want to rant about these Artificial Analysis charts that you see everywhere:
The "most attractive quadrant" is completely meaningless. The whole point of a Pareto curve is that each point on the curve is better than everything else on at least one dimension, and that you can make these comparisons without placing a value judgement on the relative importance of the different metrics. If you make a composite score of the two metrics (any monotonically non-decreasing function, e.g. a weighted sum with non-negative weights), that score will always be maximized by one of the points on the Pareto frontier.
So going by the numbers in the 2nd chart (1st AA chart) from TFA alone:
- there's no reason one would choose Deepseek V4 Pro 0813 (max) even though it's in the "most attractive quadrant", because GLM-5.3-Flash is both cheaper and scores better.
- Claude Fable 5.1 (max with fallback) on the top right* could be your most attractive option if you need the best scoring model and don't care about cost, even though it isn't in the "most attractive quadrant"
- The un-shown model off the left side of the chart could be your most attractive option if you just need lots of cheap tokens and don't care about quality.
(Obviously if you start including other factors in your score that aren't represented on the chart, then you might choose differently.)
* I also dislike the way they place the labels, and that grey line connecting the label to the point is way too subtle.
Sound critique. I'll add that the Artificial Analysis intelligence index is not considered a good metric for intelligence anymore. Most of the benchmarks that it comprises are saturated or considered low signal today.
I'm always reminded on Orwell's quote about the then new atomic bomb and his prescience on how it would all work out:
"Had the atomic bomb turned out to be something as cheap and easily manufactured as a bicycle or an alarm clock, it might well have plunged us back into barbarism, but it might, on the other hand, have meant the end of national sovereignty and of the highly-centralised police State. If, as seems to be the case, it is a rare and costly object as difficult to produce as a battleship, it is likelier to put an end to large-scale wars at the cost of prolonging indefinitely a “peace that is no peace”."
It seems, especially with open weights, that the AI is much more like the alarm clock and not the battleship. $20/mo would have been about $1 in 1944
Good point. Everyone is worried about the security attack angle of AI, but it also helps the defence side. People can use it to secure their own systems.
I've been using GLM-5.3-Flash on Ollama Cloud's $20/mo plan and using OpenCode's free models (mostly Muse Spark 1.3) to do a LOT of work and I would say that I hit a daily or weekly limit MAYBE once a month.
My usage plus reading about how tokens just keep getting cheaper and cheaper sounds like a great thing for "the rest of us" but not sure how the frontier AI labs are going to pay back all of their debt if this is the case.
(I get the inference is currently very profitable but if it's a race to the bottom on token pricing, even that won't last much longer)
A much deeper analysis on the falling price per task was published yesterday by Epoch AI [1]. It's a real statistical analysis and comes to more defensible and grounded conclusions. The headline takeaway is:
The cost of a given level of performance often falls fastest right after that level is first achieved, that is, when it is state of the art (SOTA). We see this pattern on three of our five main benchmarks of AI capability. Averaging across all five, cost falls 66% per quarter (75× per year) for performance that has just debuted as SOTA. Two years later, prices fall half as fast, at 32% per quarter (4.7× per year).
but the analysis itself has more nuance and is a quite interesting read.
It's true that LLMs "want" to be be local, but they won't shift broadly to being local until there's a sufficiently large supply of VRAM or (at least) "unified" memory from the manufacturers. (I'm also assuming here that radical regulatory changes like government bans of local models aren't going to happen.) So (AFAICS—I am no expert) the future of LLMs over the next few years comes down primarily to the nitty-gritty of how much memory fab capacity will be added and when, and to a lesser extent of what happens to future demand from LLM SaaS services (& maybe their existing stock of hardware if they get in trouble). (I'm also assuming no roughly-AGI-sized leap forward which makes the frontier models of the near future vastly more valuable than the near-fontier models of today.) For the incumbent manufacturers the high-margin business is selling to LLM SaaS providers who use VRAM efficiently, but the high-volume business is getting chips into millions of laptops which will use VRAM very inefficiently. I assume that they will want to move from high margins to high volumes as they build they physical capacity to ship higher volumes, but they seem to prefer to do it at a stately pace. Hopefully some jostling from Chinese competitors, and maybe a dropoff in demand from data centres, will speed things along.
I wouldn’t say so. Once local solutions pass a threshold of affordability consumers tend not to mind too much about their inefficient resource utilisation: see the many thousands of MacBooks which sit largely idle for most of the day. And efficient utilisation is actively against the interests of hardware manufacturers, at least while they’re not supply-constrained and looking to sell their limited supply to whomever can pay top dollar.
I think Nvidia is under the same pressure as Anthropic/OpenAI. Nvidia will dominate research and probably keep dominating training, but the real volume is in inference. And for inference Nvidia's lead is only a few months, similar to the lead frontier labs have over open source. Nvidia will sell a lot of Rubin CPX's, but their margin on that will be a lot smaller than B200 because there is so much more competition in that space.
> Nvidia will sell a lot of Rubin CPX's, but their margin on that will be a lot smaller than B200 because there is so much more competition in that space.
Given the shrinking margins, I wonder whether Nvidia will still think it's worth competing in that price-performance corner in the long term.
There seems to be a mistake in the cost comparison between 2025 and 2026. The 2025 chart axis is the cost to run the entire "intelligence index", and the 2026 version is a weighted average cost per task.
I don't disagree with the thesis here, I just don't think costs are coming down quite that quickly.
GPU case seems very weak. The graph is impossible for me to reason about at least. You could draw basically any trend line through that GPU graph and it would look equally plausible to me. The main takeaway I get is that the NVidia H100 from four whole years ago is barely different in efficiency from the state of the art, which is surprising to me, and seems to indicate the exact opposite of what the article says.
I don't know about the malleable software claim. Sure, people can build out their own thing. Malleable software in itself requires software that is designed to be malleable in the first place.
And perhaps people will not be willing to accept the initial friction that malleable software brings (see people who complain about Emacs or Salesforce or JIRA)... and end up just kinda churning indefinitely on re-implementing things over and over again.
"Internal IT teams" striking aback against SaaSes, perhaps. We'll see
The point for the cost of a task going down is true but the evidence is not correct : artificial analysis have changed their benches several times over the period shown, hence the cost as well.
for the first chart (sourced from https://epoch.ai/data/machine-learning-hardware?view=graph&y...), what is the audience supposed to think about that trend line? there's a step after you slap a regression on some points where you evaluate whether there's a real trend or noise, right? i don't see that either in the article or the linked source.
One super important thing missing: Speculative decoding. Things like Dflash(2), Dspark etc. help to do one forwards pass and get 6-7 tokens out of it. (For completeness, the embeddings from the forward pass are passed into a diffusion model which predicts the next tokens, and the model just verifies it (very cheap operation)). So we can produce way more tokens for roughly a similar amount of compute.
I agree with OP that we will continue to see improvements, but there are also some serious bottlenecks ahead of us:
- Energy is not infinite, neither energy efficiency is.
- Datacentres neither.
- Benchmarks are an abstraction of real world problems!
On top, there is an overall "economic" aspect that most of the people miss: every change carries a certain degree of risk (lose money, reputation, customers, death of people, ecc) that very few want to take and a lot of changes(e.g. rewrite some piece of SW in another Lang) don't produce a positive economic impact.
If inference continues toward the trend of becoming a commodity, and inference hardware efficiency is doubling every two years, that will likely mean that the price of inference will continue to fall so long as the market is competitive. I don't see how companies investing billions in inference hardware today will see profit in the lifespan of the hardware.
everything will be able to talk to anything else, for real this time
it will be like the internet of things only some asshole will call it "intelligence of things"
again, there will be no S for Security in this new IoT
your thermostat will one day start fucking with you. when you run a diagnostic llm on it, it turns out it's keeping around 5 different viral copies of personality files around, that were left behind by llm botnets/openclaw-like memetic replicators/your grandpa leaving behind easter eggs before his death.
the future will be pretty evenly distributed, and full of weird shit
Suddenly the ban on computer networks in Battlestar Galactica doesn't seem so far fetched! We're not that far off having enough distributed compute for LLM viruses to install corrupted LLMs everywhere and then start attacking everything within reach.
If we're all using distilled open-weight models in ASICs in our own systems the energy cost will come way down. The question is when that becomes a reasonable solution for a broad set of use cases.
I think this is a case where just drawing a "line goes up" extrapolation is incredibly misleading because there is _tremendous_ economic pressure to get costs down, and costs are very tightly tied to energy use. All of these systems are incredibly inefficient right now and have a lot of room to go down in energy use. I'd guess that the absolute _floor_ is burning model weights directly to silicon and that's like a 90+% reduction in energy use.
Not really related to the central point, by but I couldn't help but get caught up by
> Generally, models intended to be run locally will be much smaller, such as Muse Glimmer or Qwen3 Coder.
That is such an interesting set of models to use as examples here. One being essentially obsolete on release a month ago, and the other being completely ancient in LLM time. I really wonder how they landed on those two.
I just stopped reading at that, for anyone else, Please find a better source and take everything in here with a grain of salt.
IMO
The number one improvement that mattered for local AI was llama.cpp, partial offloading to system cpu/ram. The next was quants, being able to take fp16 and turn it to q8, q4 etc. The next IMHO is unsloth dynamic quant, that have been able to do mixed precision so we have UDq1/q2 that is actually pretty damn coherent. Allowing individuals to drive K3 locally even if it's at Q1/Q2. Then MoE changed everything for everyone, cloud and local. The other is integrated GPU, Apple, Strix Halo, DGX Spark. Then all the extra improvements like MTP, DSpark, etc. Of course there's many other additional things that have mattered too
They're not secretly subsidizing, they're openly subsidizing.
Token pricing was a small minority of customers up until this year, when all the labs started trying to force customers onto token-based billing. Within the last week, Anthropic repriced my team's plan from a temporary "50% extra tokens" to 25%: https://support.claude.com/en/articles/15910845-claude-code-...
The fact that all this is ongoing within such a short timeframe should make you suspicious of any analysis that claims to be observing "statistical trends" like they've discovered a new Moore's Law out of 6 months of pricing data from 2 companies.
your repricing has nothing to do with subsidising which means selling at a loss. Within this year, the real prices have gone down more than 10x on average which is way more than the teeny 25% you are fighting for.
I find it difficult to believe the inference only providers (Baseten, Fireworks, Digitalocean, etc) are all selling tokens at a loss.
Asking Claude for a rough estimate based on publicly available throughput and cost data for open weight models on modern GPUs suggests serverless, pay-as-you-go inference is profitable on owned GPUs with reasonable utilization (30-50%).
The cost is decreasing quickly, mostly because the labs stopped competing on quality. At the same time, the costs are enormous, and all labs are very openly subsidizing usage hoping to get enough scale to be profitable (while 1 of them has suspicious unity numbers and can be hiding negative marginal income).
Also, it's impossible that they become cheaper than specialized software. Or even as cheap as them. It's still possible that they become cheap enough that it doesn't matter.
how is this nonsense still so persistent? There's piles and piles of evidence that inference has massive gross margins at api pricing. what are you actually talking about?
This is the core of my belief that data center construction is a huge bubble.
AI is not a bubble, IMO, though we may see a retrench and some companies with sky-high valuations will crash to more reasonable ones. But data center demand is probably a bubble, and the main driver will be reduction in the actual amount of power and data center space required to serve escalating demand.
I think hardware and model improvements will pace or maybe outrun demand and then when demand starts to saturate will keep going and leave a lot of orphaned data centers.
I disagree. I own over 1TB of vram at home. I can tell you that it's not a bubble. From my builds, I would rather have cloud, cloud is easier. From running small models like Qwen3.8-27B to large models like Qwen3.8-2.4T. I can tell you that small models will never be enough or match up. Everyone will want the smartest model, not just a good enough model.
> Everyone will want the smartest model, not just a good enough model.
Not so sure about this. There’s always a potential threshold. After all, we don’t all use the most powerful computers, the latest phones, the highest resolution cameras, the fastest or best cars.
I am already not interested in cloud LLMs and I don’t even use the best (on paper) model that I can run locally. I prefer a model that people insisted (here) was “dead on arrival” but appears to work better for me.
Jevon's paradox isn't a physical law, it doesn't magically apply to everything. Millions more copies of Atari's ET game didn't cause everyone to pickup a cheap copy, and cause extra demand for a garbage video game. Some times (actually, usually, I'd argue) things are made that will sell for less than the cost of construction because of irrationality, and they don't induce extra demand and they don't change the negative profit margins.
You can't simply wave Jevon's paradox at things. Thousands of miles of canals were dug in the UK that couldn't be sustained and were abandoned. Thousands of miles of railways were laid that could be sustained and were abandoned. And those are potentially durable investments, unlike cheap walls, pillars and roofs laid over a levelled concrete slab full of fast depreciating IT equipment.
The article observes that the cost of frontier intelligence from 2025 has fallen 100x in the last year. It also notes that the energy to run models is also collapsing. Consumer hardware is borked right now because these new algorithms are revolutionizing the utility of a computer. Computing is technology who's cost has been collapsing for 90 years, and its a safe prediction that it will decrease again.
> Yeah okay bud, anyone checked in with the state of consumer hardware recently? Not the author, evidently.
RAM prices will crash when demand drops even a little. They'll probably crash to a lower (inflation adjusted) level than before. This has happened before.
Industrial scaling in general often looks like a sawtooth: price spike, capacity investment, crash, repeat.
Part of what's keeping prices high a little longer is that everyone knows this and is a little reluctant to plow resources into chip fabs for fear of having the bottom fall out before they recoup or sell that to someone else to hold that bag.
Graph the average compute and RAM in a mid-high end laptop at an inflation adjusted price point for the past 40 years. It's very exponential and hasn't slowed down much.
No. Prices will crash when supply side expands to meet the increased demand. Because demand won't go down to pre-bubble times any time soon. Unfortunately the supply side has been very slow in increasing production, partly because most steps of the production chain are all maxed out.
On a long enough scale you are right that prices will likely normalize to a better level, but before 2030? That would mean the factories are built quickly once they begin.
The entire reason why hardware prices are so absurd right now is because manufacturers across the board are doing everything to prevent that crash.
They all collectively chose NOT to increase supply with increased demand. So if the bubble pops, they just go back to previous prices without oversupply driving the prices to rock bottom.
> Tokens become cheaper than tool calls
The author observes that a call to GPT-5.6 Luna is only 4-5 orders of magnitude more expensive than grep, and then predicts that at current rates of progress, calling an LLM will soon be cheaper than a grep. I think this is a good time to invoke Stein's Law: "If something cannot go on forever, it will stop." These efficiency improvements won't continue forever. It's more likely that the per-call cost of high-quality, compiled software like grep will be a lower-bound that LLMs asymptotically approach, rather than a line that they blow past with perpetual exponential progress. (Barring a true breakthrough in something like quantum computing or room-temperature superconductors.)
Yeah I bumped on that too. If it's possible to make llms cheaper than current grep, then it is also almost certainly possible to make grep cheaper.
You can burn anything* into an ASIC to make it cheaper per-call.
non-backreferencing grep is not very difficult to implement in an ASIC either. But it's probably not worth it because of how relatively rarely you use it and of the data transfer costs.
LLMs are great candidates for ASIC-burning because they're slow compared even to network speeds and run all the time. The issue is that you don't want to burn a specific model or architecture that then becomes obsolete.
So you've got two possible futures, and both guarantee large price drops: (a) LLMs keep getting better and better and better, so ability/$ keeps rising; or (b) LLMs plateau in ability, in which they will start getting ASIC'd.
49 replies →
grep is deterministic. Llm is probabilistic. Llm can be transferred to a tiny quantum cpu or a lower precision float.
When the author wrote Llm can be as cheap as a tool, I read it as not equivalent. They even said the Llm can be embedded into a tool.
Their point was, the higher level use case — like classification — could become as cheap as grep. Which is quite well possible.
1 reply →
depends on what you are grepping ... greapping a large file might be more expensive one day than generating n-th token with LLM that works fully in hardware
you could make hardware implementation of grep and store the file itself next to it in some ROM but that's not a very useful grep ... while hardware LLM is exactly as useful as software LLM only orders of magnitude faster
1 reply →
4-5 orders of magnitude is huge. Assuming an order of base 10, it's 10000x-100000x. So a call to grep may return in 1s on a typical PC. That means a GPT call takes equivalent energy of 10000-100000 PCs to do the same in 1s. That's a difference that can't be equalized with scaling. It would require a revolutionary breakthrough.
I also don't understand where the idea that frontier models are getting better efficiency comes from. The results are certainly improving, but that comes from feedback and multiplexing requests, which cost more.
Recall the articles claim that efficiency is gaining 2.5 orders of magnitude per year.
Something seems off about this.
LLM is spicy memoizing, so it can potentially be faster than a tool call. But people will spend a month tweaking and testing to ensure they have the level of determinism they need, which means it's more expensive, and that they should have used actual memoization in the first place.
A classic case of someone projecting out to infinity from just after the first bend of the S curve.
> The author observes that a call to GPT-5.6 Luna is only 4-5 orders of magnitude more expensive than grep, and then predicts that at current rates of progress, calling an LLM will soon be cheaper than a grep.
At some future point where LLM hardware is cheaper than simply running grep, then grep equivalent would benefit from those selfsame hardware improvements and be cheaper to run as well, probably still by the same ratio.
Well, think that statement through a bit:
Grep reads through the entire file looking for patterns.
An LLM scans its neural net (in ways that I don't understand) which is kinda-sorta like having a huge index.
You can improve over Grep if you have an index; and the LLM has an index.
Thus, it's plausible that an LLM can be more efficient at reading its neural net (IE, index) than Grep reading the whole file.
But if the problem is literally grep (search this file you've never seen before), no index can pre-exist.
If you assume the file arrives ahead of time, can be indexed, and that this is worthwhile because we want to support multiple pattern matched retrievals, then sure it makes sense to consider indexed query schemes and upper/lower bounds. Each query could be faster as an inference if it doesn't have to re-scan the whole file.
But I don't think anybody, in good faith, can pretend that any LLM can digest a file faster than grep can. Particularly, if you admit the vector processing dedicated to doing the convolution kernel(s), you should also admit similar hardware could run a vectorized grep.
The LLM doesn't have an index of every single file I might want to grep, though. In practice it has an index of very few of them, and perhaps even none of them.
It might never beat out grep, but it could beat some more expensive to call tools, similar to how heuristics will often be faster than exact answers. Rust Analyzer can be slow at times, I could see an AI tool taking over a subset of its work.
It's probably better to optimize rust-analyzer first (and now, this became easier). I mean, see rust-glance: it's not feature complete but it points out to different tradeoffs in this space
Right. And some hardware improvements will speed up both grep and Luna, which won't close the gap.
not necessarily, one may be easier to parallelize while the other suffers some serial computation bottleneck.
3 replies →
A few years is hardly forever and the state of the world here indicates a lot of low hanging fruit still exists.
An LLM can certainly be cheaper than grep, because it’s an approximation, while a grep is deterministic and must examine every byte in what can be a relatively complex state machine for a regex based grep. There are other scales to consider like the scale of your local hardware vs the highly multitenant and high end hardware of the hyper scale inference providers.
There are already high volume models for coding inference where the reasoning time is crazy low and cheap per token where it can build reasonably simple software so blindingly fast it isn’t implausible the bottleneck is the latency in tools and networks. I find them hard to use at times because I don’t have time to think through the next turn by the time it’s done.
Regardless I wouldn’t be surprised to see a world where tokens are so cheap it’s not worth metering them but charging licensing feels with meter tiers at the far horizons to prevent abuse, charge outliers. Subscription models already set this stage well.
The other side to consider is bountiful capacity will also drive tokens to near zero price. The data center build out is barely underway and as it materializes, as hardware efficiencies improve, as techniques and model science and technology improves, harnesses, methodologies , etc improve, the economics flip from load shedding to trying to keep the data centers utilized. The economics lead to the world where tokens are not a unit of measurement for cost for anyone other than the inference providers to manage their utilization.
> An LLM can certainly be cheaper than grep, because it’s an approximation, while a grep is deterministic and must examine every byte
You mean a grep over terabytes of data vs a LLM with gigabytes of parameters?
If you have so much data, you can use an index to search. It's unlikely that LLMs are going to be cheaper than properly indexed search DBs (which is what we should be comparing them with)
There is no reason to assume that quantum computers would benefit LLMs in particular. Perhaps we could implement LLMs as analog circuits to save energy.
What if the tool is more advanced, like an optimizing compiler: `g++ -O3 -march=native -x c++ - <<EOF ... EOF`
If the compiler invocation is sufficiently slow, the llm could consider outputting a binary directly?
For all we know matrix multiplications are a faster way to generate optimized machine code than branchy sequential compiler code with tons of heuristics and passes.
> For all we know matrix multiplications are a faster way to generate
“are” or “could be”?
> calling an LLM will soon be cheaper than a grep
From a computational standpoint this is obviously nonsense, but from an attentional one I'm not so sure. It may already be more attentionally expensive to use grep in some cases, such the moment you need to remember a non standard arg. And if this applies for performing a simple http operations, then it certainly applies going up the complexity chain.
> Barring a true breakthrough in something like quantum computing or room-temperature superconductors
Won't that also help grep and then move the asymptote down more?
The cost for my employer to employ me is only 4-5 orders of magnitude over that of the snack in the vending machine…
room-temperature superconductors, sure, but I fail to see how quantum computing will disrupt – in the medium term (25 years or so) – classical computing in any meaningful way
Is running LLMs (or some other ML workload) on/with quantum computers expected to bring efficiency gains?
But grep is just a tool in a pipeline between a question in your brain and an answer you are searching for.
what if an LLM finds the answer early?
grep might continue to read everything, doing the wrong thing correctly.
I think it's an interesting thought experiment: could an LLM call be a "cheaper" grep?
Especially for deterministic activities it just feels impossible to imagine general LLM tech handling the problem better, despite everything being said.
But hey, tech is filled with "smashing the generalist hammer works better than the specialized tooling". Would be odd though!
One thing that's easy to miss about performance is that it depends on framing.
An example I like to give: optimizing a data processing program's runtime by 5x is obviously 5x speedup for everyone. But if, for some reason[0], this means it crashes and restarts more often, it stops looking like this to end users. If every restart means it needs to start from scratch, and it restarts 10 times on average now where it didn't restart before, the 5x speedup suddenly looks like 2x slowdown to end user.
In this sense, LLMs are already much more efficient than most CLI tools, by a combination of:
- User not having to remember the exact invocation, or even the name of the CLI tools
- LLM being able to run the CLI tools and chain them on its own
- LLM being able to self-correct in case it got things wrong, or when actual output show that user's idea was right, but the instructions were wrong
Prompting "okay, list those processes sorted by runtime and match them against these output files" is both faster to type than the actual commands, it also end-to-end completes much faster than doing it by hand.
--
[0] - And I don't mean a bug. Say it's some batch processing run on a cluster with aggressive resource usage management; 5x speedup means it runs much hotter than before, which may put it on the top of "kill list" for when the cluster managing code needs to free up some resources.
"Did you know that disco record sales were up 400% for the year ending 1976, if these trends continue...AY!"
AY-3-8913?
I found the OP insightful and worth a read. Thank you for sharing it on HN.
The only aspect that is poorly analyzed by the OP is business model viability. All players are investing insane amounts of money in infrastructure with the expectation that their future profits will justify all that investment. The winner or winners in the AGI race, they believe, will find the proverbial "pot of gold at the end of the rainbow."
The OP glosses over questions of business model viability with a brief qualitative discussion and very little hard data. For example, to earn an annual return > 10% on every trillion dollars of capital sunk into infrastructure, the owners of that infrastructure must earn free cash flow (operating profit less investment) in excess of $100 billion per year in perpetuity. Is that feasible? Why? How?
The OP does not really consider such questions.
I think the article's analysis is basically right in a vacuum. That is, I think it's clear that inference is a viable business model. But what isn't clear is whether it will be such a profitable business model for any given company that it will justify the investment that company has taken. I kind of think the winners might be a follow-on generation of companies that focus on this commodity inference business model instead of the invent-machine-god-first "business model" and thus are wiser about their level of investment and capital costs.
> I think it's clear that inference is a viable business model.
You may be right. I'm not so sure. Inference looks like a viable business model for those operators that have SOTA infrastructure in place, but the investment required to have it is enormous, and appears to be never-ending, because if an operator stops investing aggressively, its infrastructure quickly becomes non-competitive, and customers will quickly leave for alternatives. SOTA infrastructure is a moving target.
5 replies →
> I think it's clear that inference is a viable business model.
Only if you also have the model thats better than anyone else's.
As soon as models are free, or there are no newer models (assuming thats going to happen, and thats not a given) then the only thing you can compete on is price.
This means that the only thing you have to differentiate is either price, speed or ease of use. (or regulatory capture...)
We are at pets.com level of spend currently. Unless model development becomes cheaper, then we are going to run out of novel debt but not really debt mechanisms.
1 reply →
A phrase comes to mind: "Your margin is my opportunity."
There's some alternative analysis here that the current economics aren't too bad https://x.com/ramez/status/2102506182279573962
Also another take on AI costs falling https://x.com/EpochAIResearch/status/2102510281176023529 Methodology is how much it costs to do the same task a while later. Gives fallen ~47% a quarter.
The problem source is that the "cost" of tokens are taken at face value from business that are losing money at record speeds. E.g. https://artificialanalysis.ai says "doing task A costed us $10 using OpenAI", and that is the "cost" the OP used as basis for "tokens are cheap". Meanwhile OpenAI is losing $19 for each $1 in revenue... So right now OpenAI should be charging around $200 to do task A just to break even, but that would mean their use base would collapse.
There are two different things occurring here.
One is how much does it cost OpenAI to train the model.
The other is, if I stole OpenAI's model how much would it cost for me to run it?
R&D costs versus operational costs. Operational costs are very likely profitable. R&D is catastrophically expensive currently.
1 reply →
>The winner or winners in the AGI race, they believe, will find the proverbial "pot of gold at the end of the rainbow."
The hubris of this is really astounding too. There is no technology out there that some company develops and has not been reverse engineered and copied and manufactured at scale by competitors before long. You can't stop this from happening. People will leave the company or be poached and proliferate what they have built in the past. Every country that wanted a nuke has a nuke, after all.
The only way to keep the secret fire from leaking out would be to have AGI's first move be to lock the doors and prevent anyone from ever leaving company property again.
Google's search has had a long run, and seems much less complex than LLMs.
1 reply →
They are already turning profits and inference has shown to be a cash cow. And they've already secured compute for the next several years.
Some frontier labs are reporting positive "adjusted EBITDA" (earnings before interest, taxes, depreciation, and amortization, with extra adjustments to make the figure positive).
Free cash flow (operating profit less investment), actual cash coming in, is deeply in the red.
EBITDA can be a sensible measure of profitability when there isn't much need for additional investment. That doesn't seem to be the case with these operators. They need to invest aggressively to avoid losing customers to competitors. All of these operators have made multi-year commitments to invest more in infrastructure. In addition, they have guaranteed quite a bit of debt to fund it.
Maybe it all will work out fine (and I sure hope it does!), but I didn't see any hard data from the OP, or from you, supporting that view.
1 reply →
Definitely. The question is: Is it enough to recoup the enormous capital costs and justify the level of investment they've received. I think there's a decent chance that it will be. But maybe not. And the longer they keep focusing on training new models more so than on inference, the more uncertain I become that it's all going to work out.
1 reply →
Labs are playing money games with EBITDA, which is not uncommon, but also hides the extent to which they are in the red (deeply, deeply, in the red, and projected by them to get worse).
Who is the "they" that are turning profits?
1 reply →
Think of it like spending $1 trillion to become the next Google. That hardware itself may never turn a profit. But if 10 years for now you're the software provider that owns the ecosystem around too cheap to meter tokens you've got a money printing machine.
Exactly. "Cost-to-distill" is a critical parameter. Right now usage of frontier models for all tasks is both subsidized and irrationally popular even at the subsidized price. Deepseek would solve most tasks faster and 10x cheaper. I agree with the author that just as Deloitte exists, frontier labs will exist. But not because their products are proprietary technical marvels or gods, but rather because of branding.
DS wouldn't be 10x cheaper than the subsidized subscription plans from openai/anthropic. Although it is of course much cheaper than the enterprier/API pricing- I think if you're on the subscription plans, you can't beat that on performance per price.
1 reply →
Too Cheap to Meter reminds me of the promise of Nuclear Power in 1954
"It is not too much to expect that our children will enjoy in their homes electrical energy too cheap to meter,..." Lewis Strauss
https://en.wikipedia.org/wiki/Too_cheap_to_meter#Origins
Oddly enough my power bill was metered and big.
It's a deliberate reference/meme that is basically used to acknowledge the precedent of overly exuberant predictions of cost in an emerging technology but argue "however, this time it's true".
Of course, perilous territory for future irony depending on how your prediction plays out.
I remember up to around the time the PC came out, you were billed for computer time.
Later, you were billed for time connected to the "internet" (compuserve or aol or whatever)
Around when the iphone came out, software went from tens or hundreds of dollars to pennies, then free.
On the other hand legal advice has always been expensive, because a good answer is worth it.
Medical advice is worth it. Investing advice is worth it.
(That said, I wonder if with home solar and batteries if electricity will ever "generally" go down in price to normal people)
1 reply →
It's really quite unfortunate that the promise was not delivered, mostly for political reasons. I hope that a new wave of reactors and the dire need for clean energy restarts the nuclear race.
The economics are not there. Solar + batteries are much cheaper per watt today and still improving. Even better, the solar can come online instantly and expand while US nuclear takes twenty years to start generating any energy.
13 replies →
One excellent nerd-sniping side-effect of those political reasons is that the good folks of Austria built an entire nuclear power station and then never put it into service [0]. Open for guided tours on Fridays [1].
[0] https://en.wikipedia.org/wiki/Zwentendorf_Nuclear_Power_Plan...
[1] https://windows2008.zwentendorf.com/en/location-npp/
Nuclear was never going to give us too cheap to meter regardless of politics. Uranium just isn’t that cheap. Fuel costs are lower then coal or gas, but not so low that operators would just not bother to charge for it
3 replies →
Having your own solar can be "too cheap to meter", though it is definitely a feast and a famine kind of a thing.
We had batteries full and were spending electricity on all kinds of luxury things like whole-day internet and desalination for several weeks, and now that it has rained for over a week we're starting to turn non-critical systems off to keep the lights on.
On the other hand, this did work out in other areas. I pay a flat monthly rate for all-I-care-to-eat internet access, for example. My email provider has limits on storage space but I don’t get charged per email sent or received. It’s not a crazy concept on its face, nuclear power just didn’t work out as well as was hoped.
https://www.wired.com/2008/02/ff-free/
"Between digital economics and the wholesale embrace of King's Gillette's experiment in price shifting, we are entering an era when free will be seen as the norm, not an anomaly. How big a deal is that? Well, consider this analogy: In 1954, at the dawn of nuclear power, Lewis Strauss, head of the Atomic Energy Commission, promised that we were entering an age when electricity would be "too cheap to meter." Needless to say, that didn't happen, mostly because the risks of nuclear energy hugely increased its costs. But what if he'd been right? What if electricity had in fact become virtually free? The answer is that everything electricity touched—which is to say just about everything—would have been transformed. Rather than balance electricity against other energy sources, we'd use electricity for as many things as we could—we'd waste it, in fact, because it would be too cheap to worry about."
... What Mead understood is that a psychological switch should flip as things head toward zero. Even though they may never become entirely free, as the price drops there is great advantage to be had in treating them as if they were free. Not too cheap to meter, as Atomic Energy Commission chief Lewis Strauss said in a different context, but too cheap to matter. Indeed, the history of technological innovation has been marked by people spotting such price and performance trends and getting ahead of them."
My issue is that datacenter energy costs are being prioritized for commerce over residential use, so the average consumer is paying more for electricity, because a datacenter needs more electricity and they're getting tax breaks. Assuming this all improves efficiency for new products like automated robotics, there is a debatable benefit. Jevon's Paradox has no ceiling, except the environment, and people's 401ks.
I just want to rant about these Artificial Analysis charts that you see everywhere:
The "most attractive quadrant" is completely meaningless. The whole point of a Pareto curve is that each point on the curve is better than everything else on at least one dimension, and that you can make these comparisons without placing a value judgement on the relative importance of the different metrics. If you make a composite score of the two metrics (any monotonically non-decreasing function, e.g. a weighted sum with non-negative weights), that score will always be maximized by one of the points on the Pareto frontier.
So going by the numbers in the 2nd chart (1st AA chart) from TFA alone:
(Obviously if you start including other factors in your score that aren't represented on the chart, then you might choose differently.)
* I also dislike the way they place the labels, and that grey line connecting the label to the point is way too subtle.
Sound critique. I'll add that the Artificial Analysis intelligence index is not considered a good metric for intelligence anymore. Most of the benchmarks that it comprises are saturated or considered low signal today.
Would it be better to have a shaded region parallel to the Pareto curve that gets darker away from it that is labeled "better value"?
I'm always reminded on Orwell's quote about the then new atomic bomb and his prescience on how it would all work out:
"Had the atomic bomb turned out to be something as cheap and easily manufactured as a bicycle or an alarm clock, it might well have plunged us back into barbarism, but it might, on the other hand, have meant the end of national sovereignty and of the highly-centralised police State. If, as seems to be the case, it is a rare and costly object as difficult to produce as a battleship, it is likelier to put an end to large-scale wars at the cost of prolonging indefinitely a “peace that is no peace”."
It seems, especially with open weights, that the AI is much more like the alarm clock and not the battleship. $20/mo would have been about $1 in 1944
https://www.orwellfoundation.com/the-orwell-foundation/orwel...
Good point. Everyone is worried about the security attack angle of AI, but it also helps the defence side. People can use it to secure their own systems.
I've been using GLM-5.3-Flash on Ollama Cloud's $20/mo plan and using OpenCode's free models (mostly Muse Spark 1.3) to do a LOT of work and I would say that I hit a daily or weekly limit MAYBE once a month.
My usage plus reading about how tokens just keep getting cheaper and cheaper sounds like a great thing for "the rest of us" but not sure how the frontier AI labs are going to pay back all of their debt if this is the case.
(I get the inference is currently very profitable but if it's a race to the bottom on token pricing, even that won't last much longer)
A much deeper analysis on the falling price per task was published yesterday by Epoch AI [1]. It's a real statistical analysis and comes to more defensible and grounded conclusions. The headline takeaway is:
The cost of a given level of performance often falls fastest right after that level is first achieved, that is, when it is state of the art (SOTA). We see this pattern on three of our five main benchmarks of AI capability. Averaging across all five, cost falls 66% per quarter (75× per year) for performance that has just debuted as SOTA. Two years later, prices fall half as fast, at 32% per quarter (4.7× per year).
but the analysis itself has more nuance and is a quite interesting read.
[1] https://epoch.ai/publications/the-plunging-price-of-thought
It's true that LLMs "want" to be be local, but they won't shift broadly to being local until there's a sufficiently large supply of VRAM or (at least) "unified" memory from the manufacturers. (I'm also assuming here that radical regulatory changes like government bans of local models aren't going to happen.) So (AFAICS—I am no expert) the future of LLMs over the next few years comes down primarily to the nitty-gritty of how much memory fab capacity will be added and when, and to a lesser extent of what happens to future demand from LLM SaaS services (& maybe their existing stock of hardware if they get in trouble). (I'm also assuming no roughly-AGI-sized leap forward which makes the frontier models of the near future vastly more valuable than the near-fontier models of today.) For the incumbent manufacturers the high-margin business is selling to LLM SaaS providers who use VRAM efficiently, but the high-volume business is getting chips into millions of laptops which will use VRAM very inefficiently. I assume that they will want to move from high margins to high volumes as they build they physical capacity to ship higher volumes, but they seem to prefer to do it at a stately pace. Hopefully some jostling from Chinese competitors, and maybe a dropoff in demand from data centres, will speed things along.
LLM hardware wants to be shared. It's significantly more efficient economically to have expensive hardware be better utilized.
I wouldn’t say so. Once local solutions pass a threshold of affordability consumers tend not to mind too much about their inefficient resource utilisation: see the many thousands of MacBooks which sit largely idle for most of the day. And efficient utilisation is actively against the interests of hardware manufacturers, at least while they’re not supply-constrained and looking to sell their limited supply to whomever can pay top dollar.
> NVIDIA will still boom
I think Nvidia is under the same pressure as Anthropic/OpenAI. Nvidia will dominate research and probably keep dominating training, but the real volume is in inference. And for inference Nvidia's lead is only a few months, similar to the lead frontier labs have over open source. Nvidia will sell a lot of Rubin CPX's, but their margin on that will be a lot smaller than B200 because there is so much more competition in that space.
> Nvidia will sell a lot of Rubin CPX's, but their margin on that will be a lot smaller than B200 because there is so much more competition in that space.
Given the shrinking margins, I wonder whether Nvidia will still think it's worth competing in that price-performance corner in the long term.
There seems to be a mistake in the cost comparison between 2025 and 2026. The 2025 chart axis is the cost to run the entire "intelligence index", and the 2026 version is a weighted average cost per task.
I don't disagree with the thesis here, I just don't think costs are coming down quite that quickly.
GPU case seems very weak. The graph is impossible for me to reason about at least. You could draw basically any trend line through that GPU graph and it would look equally plausible to me. The main takeaway I get is that the NVidia H100 from four whole years ago is barely different in efficiency from the state of the art, which is surprising to me, and seems to indicate the exact opposite of what the article says.
I don't know about the malleable software claim. Sure, people can build out their own thing. Malleable software in itself requires software that is designed to be malleable in the first place.
And perhaps people will not be willing to accept the initial friction that malleable software brings (see people who complain about Emacs or Salesforce or JIRA)... and end up just kinda churning indefinitely on re-implementing things over and over again.
"Internal IT teams" striking aback against SaaSes, perhaps. We'll see
The point for the cost of a task going down is true but the evidence is not correct : artificial analysis have changed their benches several times over the period shown, hence the cost as well.
for the first chart (sourced from https://epoch.ai/data/machine-learning-hardware?view=graph&y...), what is the audience supposed to think about that trend line? there's a step after you slap a regression on some points where you evaluate whether there's a real trend or noise, right? i don't see that either in the article or the linked source.
My biggest takeaway is Laya and that Jev isn't novel. I'll look into finetuning Laya.
One super important thing missing: Speculative decoding. Things like Dflash(2), Dspark etc. help to do one forwards pass and get 6-7 tokens out of it. (For completeness, the embeddings from the forward pass are passed into a diffusion model which predicts the next tokens, and the model just verifies it (very cheap operation)). So we can produce way more tokens for roughly a similar amount of compute.
I agree with OP that we will continue to see improvements, but there are also some serious bottlenecks ahead of us:
- Energy is not infinite, neither energy efficiency is. - Datacentres neither. - Benchmarks are an abstraction of real world problems!
On top, there is an overall "economic" aspect that most of the people miss: every change carries a certain degree of risk (lose money, reputation, customers, death of people, ecc) that very few want to take and a lot of changes(e.g. rewrite some piece of SW in another Lang) don't produce a positive economic impact.
If inference continues toward the trend of becoming a commodity, and inference hardware efficiency is doubling every two years, that will likely mean that the price of inference will continue to fall so long as the market is competitive. I don't see how companies investing billions in inference hardware today will see profit in the lifespan of the hardware.
jgrep is not meant to replace grep. It’s supposed to extend it to new use cases. Keyword search will always have a place!
> We are likely to see LLMs integrated into every part of computing as infrastructure, not just as a product, in the next year or two.
Of course, for collecting better telemetry using local AI for analyzing video from camera and audio from a microphone.
If you can solve the problem without an AI - do that.
I'll be back next year reminding you about your cybersecurity claims.
What does this mean for the people who invested trillions of dollars in these companies?
everything will have llms
everything will be able to talk to anything else, for real this time
it will be like the internet of things only some asshole will call it "intelligence of things"
again, there will be no S for Security in this new IoT
your thermostat will one day start fucking with you. when you run a diagnostic llm on it, it turns out it's keeping around 5 different viral copies of personality files around, that were left behind by llm botnets/openclaw-like memetic replicators/your grandpa leaving behind easter eggs before his death.
the future will be pretty evenly distributed, and full of weird shit
Suddenly the ban on computer networks in Battlestar Galactica doesn't seem so far fetched! We're not that far off having enough distributed compute for LLM viruses to install corrupted LLMs everywhere and then start attacking everything within reach.
See also: Cyberpunk's Blackwall
It is hard to see how the environmental side effects of this aren't going to be somewhere between bad and disastrous.
If we're all using distilled open-weight models in ASICs in our own systems the energy cost will come way down. The question is when that becomes a reasonable solution for a broad set of use cases.
I think this is a case where just drawing a "line goes up" extrapolation is incredibly misleading because there is _tremendous_ economic pressure to get costs down, and costs are very tightly tied to energy use. All of these systems are incredibly inefficient right now and have a lot of room to go down in energy use. I'd guess that the absolute _floor_ is burning model weights directly to silicon and that's like a 90+% reduction in energy use.
No it'll be fine as long as you do your part and not drive a car, or have AC, or eat meat, or have children, or live in detached housing, or...
Magical thinking intensifies.
What should I do with this old Juicero that's taking up space on my kitchen counter?
Not really related to the central point, by but I couldn't help but get caught up by
> Generally, models intended to be run locally will be much smaller, such as Muse Glimmer or Qwen3 Coder.
That is such an interesting set of models to use as examples here. One being essentially obsolete on release a month ago, and the other being completely ancient in LLM time. I really wonder how they landed on those two.
Improvements that affect local AI - Mamba...
I just stopped reading at that, for anyone else, Please find a better source and take everything in here with a grain of salt.
IMO
The number one improvement that mattered for local AI was llama.cpp, partial offloading to system cpu/ram. The next was quants, being able to take fp16 and turn it to q8, q4 etc. The next IMHO is unsloth dynamic quant, that have been able to do mixed precision so we have UDq1/q2 that is actually pretty damn coherent. Allowing individuals to drive K3 locally even if it's at Q1/Q2. Then MoE changed everything for everyone, cloud and local. The other is integrated GPU, Apple, Strix Halo, DGX Spark. Then all the extra improvements like MTP, DSpark, etc. Of course there's many other additional things that have mattered too
We have people suggesting that ai is so costly to run that all labs are secretly subsidising tokens and we can expect a reprice soon.
Then we have these articles that say tokens will get so cheap that labs won’t know how to make profit.
Who is correct?
They're not secretly subsidizing, they're openly subsidizing.
Token pricing was a small minority of customers up until this year, when all the labs started trying to force customers onto token-based billing. Within the last week, Anthropic repriced my team's plan from a temporary "50% extra tokens" to 25%: https://support.claude.com/en/articles/15910845-claude-code-...
The fact that all this is ongoing within such a short timeframe should make you suspicious of any analysis that claims to be observing "statistical trends" like they've discovered a new Moore's Law out of 6 months of pricing data from 2 companies.
your repricing has nothing to do with subsidising which means selling at a loss. Within this year, the real prices have gone down more than 10x on average which is way more than the teeny 25% you are fighting for.
I find it difficult to believe the inference only providers (Baseten, Fireworks, Digitalocean, etc) are all selling tokens at a loss.
Asking Claude for a rough estimate based on publicly available throughput and cost data for open weight models on modern GPUs suggests serverless, pay-as-you-go inference is profitable on owned GPUs with reasonable utilization (30-50%).
The cost is decreasing quickly, mostly because the labs stopped competing on quality. At the same time, the costs are enormous, and all labs are very openly subsidizing usage hoping to get enough scale to be profitable (while 1 of them has suspicious unity numbers and can be hiding negative marginal income).
Also, it's impossible that they become cheaper than specialized software. Or even as cheap as them. It's still possible that they become cheap enough that it doesn't matter.
how is this nonsense still so persistent? There's piles and piles of evidence that inference has massive gross margins at api pricing. what are you actually talking about?
1 reply →
The labs themselves when they openly say that they’re subsidizing tokens, I’d imagine.
where? I mean the API costs
This is the core of my belief that data center construction is a huge bubble.
AI is not a bubble, IMO, though we may see a retrench and some companies with sky-high valuations will crash to more reasonable ones. But data center demand is probably a bubble, and the main driver will be reduction in the actual amount of power and data center space required to serve escalating demand.
I think hardware and model improvements will pace or maybe outrun demand and then when demand starts to saturate will keep going and leave a lot of orphaned data centers.
> AI is not a bubble
When people say "AI is a bubble", they mean economically as a whole, which includes data centers.
Perhaps we need better terminology for "product useful; numbers nonsensical"
I disagree. I own over 1TB of vram at home. I can tell you that it's not a bubble. From my builds, I would rather have cloud, cloud is easier. From running small models like Qwen3.8-27B to large models like Qwen3.8-2.4T. I can tell you that small models will never be enough or match up. Everyone will want the smartest model, not just a good enough model.
> Everyone will want the smartest model, not just a good enough model.
Not so sure about this. There’s always a potential threshold. After all, we don’t all use the most powerful computers, the latest phones, the highest resolution cameras, the fastest or best cars.
I am already not interested in cloud LLMs and I don’t even use the best (on paper) model that I can run locally. I prefer a model that people insisted (here) was “dead on arrival” but appears to work better for me.
5 replies →
Jevon's paradox says that if data centers can serve a lot more tokens per dollar or watt there will be increased demand for data centers.
Jevon's paradox isn't a physical law, it doesn't magically apply to everything. Millions more copies of Atari's ET game didn't cause everyone to pickup a cheap copy, and cause extra demand for a garbage video game. Some times (actually, usually, I'd argue) things are made that will sell for less than the cost of construction because of irrationality, and they don't induce extra demand and they don't change the negative profit margins.
You can't simply wave Jevon's paradox at things. Thousands of miles of canals were dug in the UK that couldn't be sustained and were abandoned. Thousands of miles of railways were laid that could be sustained and were abandoned. And those are potentially durable investments, unlike cheap walls, pillars and roofs laid over a levelled concrete slab full of fast depreciating IT equipment.
6 replies →
[flagged]
[flagged]
[flagged]
[flagged]
[dead]
[dead]
[flagged]
Enshittification maxxing /s
[dead]
>We are likely to see LLMs running locally at current frontier-quality on commodity hardware in the next 3-6 years
Yeah okay bud, anyone checked in with the state of consumer hardware recently? Not the author, evidently.
>oh in 3-6 years this will all be over
Yeah I'm sure Samsung, Nvidia and sk hynix will all be very calm with lower volumes and lower margins.
Seems like 3-6 years is a pretty conservative timeline to me. Memory and inference chip production could ramp up massively in that amount of time.
The article observes that the cost of frontier intelligence from 2025 has fallen 100x in the last year. It also notes that the energy to run models is also collapsing. Consumer hardware is borked right now because these new algorithms are revolutionizing the utility of a computer. Computing is technology who's cost has been collapsing for 90 years, and its a safe prediction that it will decrease again.
> Yeah okay bud, anyone checked in with the state of consumer hardware recently? Not the author, evidently.
RAM prices will crash when demand drops even a little. They'll probably crash to a lower (inflation adjusted) level than before. This has happened before.
Industrial scaling in general often looks like a sawtooth: price spike, capacity investment, crash, repeat.
Part of what's keeping prices high a little longer is that everyone knows this and is a little reluctant to plow resources into chip fabs for fear of having the bottom fall out before they recoup or sell that to someone else to hold that bag.
Graph the average compute and RAM in a mid-high end laptop at an inflation adjusted price point for the past 40 years. It's very exponential and hasn't slowed down much.
No. Prices will crash when supply side expands to meet the increased demand. Because demand won't go down to pre-bubble times any time soon. Unfortunately the supply side has been very slow in increasing production, partly because most steps of the production chain are all maxed out.
On a long enough scale you are right that prices will likely normalize to a better level, but before 2030? That would mean the factories are built quickly once they begin.
> everyone knows this and is a little reluctant to plow resources into chip fabs
except cxmt who is plowing resources in like crazy
1 reply →
The entire reason why hardware prices are so absurd right now is because manufacturers across the board are doing everything to prevent that crash.
They all collectively chose NOT to increase supply with increased demand. So if the bubble pops, they just go back to previous prices without oversupply driving the prices to rock bottom.
So this week's coordinated shock and awe campaign is about "open models" and "prices are falling rapidly". Also on the front page:
https://news.ycombinator.com/item?id=49815526