Comment by spott
12 hours ago
Yea, but they would have to become insolvent in a way that makes compute lose value.
The reason Nvidia is comfortable making these deals is because if OpenAI can’t use the compute, someone else can.
Granted OpenAI going insolvent likely means a drop in the value of compute…
Compute has already lost value for me. Six months ago I thought you needed a 1T+ model to be useful coding. Now I am able to get by just fine with a 27b model.
I see two factors converging to cause a collapse of this house of cards:
1. People are realizing that what they need isn't more general intelligence, it's more specialization. A small but well tuned coding model, a small but well tuned customer service model, a small but well tuned document explorer.
2. Specialized hardware - TPUs and NPUs - especially coming out of china. The latest GLM model was trained and runs on Huawei hardware. Nvidia is only worth so much because they are the biggest and best provider of the kind of compute needed to run llms, but the export bans mean china has a lot of incentive to topple that monopoly.
The amount of compute we need to do the things llms do is falling rapidly, the number of people who can provide that compute is rising.
> People are realizing that what they need isn't more general intelligence, it's more specialization. A small but well tuned coding model...
It’s not quite as simple as that. Several studies have shown the opposite: models trained on more diverse knowledge tend to cross-pollinate across domains. So a more generalized model can actually perform better than a specialized one.
That’s why you’re not seeing tons of tiny models (one for Python, one for Pascal, one for Rust, etc).
This is definitely the position of the big ai companies.
But it doesn't match my experience. Qwen3.8 27b is clearly smarter at coding than MANY bigger models. gpt-oss-120b for example, is almost 4x the size, and performs way worse at coding tasks.
It's clear to me that you can build small models that work well at specific tasks.
Python vs Rust is probably too fine grained a way to build a model. Coding in general seems like a better target.
There will always be a place for large generalist models, no doubt. But I think that place is much smaller than the big ai companies are counting on.
12 replies →
The western labs are very AGI pilled, and their public models are distilled down from larger research-only models that are uneconomical to serve directly. They could (and probably will) start distilling models for more niche use cases eventually, but we're not there yet.
> more diverse knowledge tend to cross-pollinate across domains
Yeah, the cross domain transfer learning from RL is overstated by a lot.
Problem is conflict of interest: the studies are mostly from the providers of the biggest models, or someone who received free tokens to do the research.
5 replies →
I think what will keep the industry afloat, all else failing, is the surveillance industry! Nothing like a fat reoccurring cheque from the government to check if little Jimmy is committing thought crime!
LLMs needing less compute would actually be a good thing for Nvidia due to Jevons paradox. Right now token costs are an impediment to using AI more broadly, and more efficient models would help adoption in cases where AI has proven to be useful, like coding.
https://en.wikipedia.org/wiki/Jevons_paradox
Jevon’s paradox is a common talking point but it is not a law of nature. LED lightbulbs use 80% less energy than incandescent but you don’t see people using 5x more lights on their homes. The overall energy used to light homes has decreased.
And even if compute demand were perfectly elastic it’s only a good thing insofar as it drives demand for new Nvidia hardware. If tokens can be served from Apple hardware or Google hardware or Huawei hardware that doesn’t help Nvidia.
2 replies →
Perhaps. But their huge valuation is based on them supplying the massive buildout of data centers that’s happening / planned.
If that dies because a lot of people’s needs turn out to be met by a system at home they can run a 30b-150b model on, a lot more of that money goes to apple or intel or amd.
That's unless the code produced in the future is much more complex than today's.
Sure, but it would be actively bad to make the code more complex simply because we have machinery that helps us deal with the complexity. A big part of how people assess the models' coding capability is whether they create needless, incidental complexity.
1 reply →
Assuming it’s all going to be vibe coded garbage, yeah it will be much more complex. Like a toddler writing a symphony.
You're not considering video which OpenAI opted out of when they retired Sora.
Generative video requires significantly more computing power and energy than generative text.
OpenAI is fucked, compute is still needed, it's just them that isn't.
OpenAI dropped sora because it was costing them ridiculous amounts of money and earning them very little. They determined that the market can't support the cost of generating video.
Without a material change in the market (more buyers, vastly cheaper generation), it's unlikely a different company could make that work. More buyers isn't likely to happen, so that leaves vastly cheaper generation - something that would cause nvidia's value to collapse if it happened.
3 replies →
Minimax H3 works pretty great and you can run it on a 3090.
oAI isn't anywhere near close to fucked as long as their models are head and shoulders above even the very best open models in terms of tool calling and rock solid stability/reliability for agents/coding harnesses. Which, they are right now and we'll see if open models actually catch up in that regard. Even the "best" open models pale in comparison with tool calling and general "prompt and go do something else for an hour" reliability that we have with GPT models. With GPT models, streaming rarely stops unexpectedly. You almost never have to constantly nudge them along, etc. Granted with open models all of this can vary depending on the provider, and perhaps open models/protocols/APIs/harnesses aren't well enough aligned, but OpenAI models just seem to work without constant (or hardly any) wrinkles and with almost any harness/agent.
4 replies →
There would also need to exist sufficient demand for video, which hasn’t happened yet.
I've been thinking about that and that's why Nvidia's prices are surprising to me. Investors should know that better than me so there must be something I don't know
It’s really hard to know when the large tech companies have so many shares owned by a single figure. They can use margin loans and options to create the appearance of demand.
This is the right kind of analysis, but we can look broader. Both the demand and supply situations are a lot more extreme and dynamic than appears at first glance. E.g. to your points:
1. Yes, smaller models will become more popular, especially as the tokenmaxxing trend dies down and people start stretching their budgets farther. That is a downward pressure on demand.
But along the same dimension, consider that currently only about 40 - 60% of the world uses AI for only about 5 - 15% of their work hours. That means there is still 2x growth from users and 7x - 20x growth from the rest of the work hours left to capture! That is 14x - 40x more demand. Then consider that agentic tasks require multiples more tokens, and that is the kind of usage that is most likely to be deployed, and also the kind of usage that is the least used right now. That's another huge multiple to be tacked on.
And the entire AI industry has been lamenting the extreme compute crunch they're facing (and also why Claude has 9's comparable to GitHub; whereas OpenAI has been chugging along because Altman was OK being called a "podcasting bro" while desperately scrounging for compute years in advance.)
Nvidia's meteoric rise is entirely due to this kind of exploding demand with extremely limited supply.
2. Competing hardware is definitely a threat, but it has its own hurdles. Because the real bottleneck is not Nvidia, it's TSMC.
Pretty much all demand for all chips in all devices in all the world flow to, like, 3 companies in the world that actually fabricate them, and TSMC is the biggest. And the supply is extremely tight, as the exploding costs of electronics clearly shows.
So now TSMC will of course try to keep all its customers happy, but it will inevitably be forced to choose which ones it will keep happiest. And those will be the customers who can pay it the most. And that would be the one with all the money from its de facto status as a monopoly (and possibly even a monopsony)...
Which would be Nvidia ;-)
So yes, compute per task is falling rapidly... but it's barely a dent in the humongous total addressable demand, and the amount of hardware to support that compute is still very constrained, and most of that supply will likely flow through Nvidia.
> But along the same dimension, consider that currently only about 40 - 60% of the world uses AI for only about 5 - 15% of their work hours.
Ah yes, i am constantly lamenting that my barista isn’t using ai enough ;)
Hopefully you’ve adjusted your ceiling numbers to account for the large amount of people who can’t afford to pay for llms, and will never be able to pay, and aren’t worth it to advertise to since they can afford very little
They need the right harness and either your help it auto produces in time enough content to further improve.
A classic big ai talking point. Color me skeptical.
If you reshuffle your argument, and apply the same facts you get to a similar conclusion but with a drastically different spin.
> it's more specialization
China, constrained by hardware, and talent (not to slight the Chinese, but they are limited to domestic resources - and much of the US effort is very international). They did, what the Chinese do, and optimized the process of production, and drastically lowered the cost of development of their models. Cheeper to build, cheaper to run is just good economics.
Meanwhile in the us, we have open AI doing "experiments" - it looks like the costs around the hugging face hack are going to be about the same as China would spend on building out one of their smaller efforts (several million dollars). (Depending on whos numbers you trust, the fact that I can even make this claim should make you raise an eyebrow).
Go back to the 80s' and "expert systems" - most people will tell you that for their time, they were amazing, and useful. People would have loved to have more of them but they were so cost prohibitive that we all but abandoned them for serious use. The US frontier labs seem to have forgotten this lesson and their calls to "slow down" look like an excuse to "cut the waste so we can move to making money".
> if OpenAI can’t use the compute, someone else can.
The problem with that is that OpenAI can only afford to pay for the compute because they are burning investor money (and so are most of OpenAI's biggest clients). They are losing billions. If they stop burning money, nobody else will be there to pay for that compute at OpenAI's cost.
Sure, somebody will probably be able to use these GPUs, they just won't be able to pay nearly as much for them as OpenAI does.
In reality, it's just nowhere near worth as much as OpenAI pays for it. Inflating the cost of compute is part of the problem caused by the circular financing, and if (or maybe when) OpenAI goes, the price of compute will go with them.
> investor money
But that’s the point. Investors believe investment in AI will pay off.
Right, "believe".
4 replies →
Devil's advocate: OpenAI not being able to use compute is highly correlated to many other AI companies not being able to find a meaningful use of this compute.
Failure to take into consideration those kind of correlations ("If my biggest client isn't able to buy it, I would be able to find someone else who will") is one of the principle causes why many risk models turned out to be garbage during the Great Financial Crisis.
That is why I added the last line.
But I also doubt Nvidia is on the hook if OpenAI just no longer wants the compute. I bet they are only on the hook if OpenAI cannot pay for it (is insolvent in some way).
I also have to bring up that OpenAI has already spat out an inference chip that beats Nvidia on flops per watt. So they could potentially not need the compute while other ai companies do.
The problem is that if OpenAI doesn't want the compute nobody does. All of these companies' demand for compute are correlated. It isn't likely that OpenAI will want less compute in isolation. Furthermore, the circular financing structure means that if OpenAI buys less chips it means that Nvidia has less money to give to say Anthropic to buy more chips and suddenly the exponential growth that circular financing has allowed to grow runs in reverse.
> they would have to become insolvent in a way that makes compute lose value
They would have to go insolvent in a way that hits Nvidia revenue. Those are related by distinct factors, a difference that may matter in a crisis.
Maybe I’m just old and cynical but that sounds like the kind of assumption of independence of events that led to the GFC .
If AI wasn’t a thing would Nvidia be worth 1/10th is current value?
Gaming is lucrative but not THAT lucrative…
Granted OpenAI going insolvent
All the "frontier" AI companies *are* currently insolvent. They have never been anything other than cash burning machines.
The only way they keep the lights on and the doors open is by borrowing money --- and epic amounts of it. If those operating the cash spigot decide to turn it off, all AI companies will likely be similarly affected --- and so will Nvidia.
OpenAI expects to burn through more cash between 2024 and 2029 than Uber, Tesla, Amazon and Spotify did - combined - before those companies started making money
https://www.morningstar.com/news/marketwatch/20251205243/thi...
To make things worse, hardware prices have spiked, due to AI companies.
Fairly sure data center construction costs are also going up (they require so many resources that everything is constrained at the moment, especially electricity production).
So I don't understand in what world these frontier AI companies can somehow become profitable. The basic tech they're using is basically the same. Yes, around the edges there are a lot of things that can be done, and were done, like caching, batching, mixture of experts, etc, but basically everyone has done all of that by now, and they're still losing money.
So:
Total costs going up a lot - revenues per unit not increasing proportionally, if anything, Chinese models are forcing those down.
How does that math work out to profits? I don't see it.
Or about as bad, after trillions of dollars in investments over multiple years, let's say the entire frontier AI sector has a total profit of $20bn by 2030. In what world does that make sense? Assuming they can scale that total profit to $100bn in 2035 without investing another cent from 2027 to 2035 (utterly ridiculous), the return on investment would happen in roughly 20 years.
It's capitalism run amuck --- and on an epic scale.
China is the one that is really in the driver's seat here. They have the opportunity and the ability to nullify/wipe out our huge investment in AI.
> if OpenAI can’t use the compute, someone else can
This is the big point IMO since I have never given $1 to OpenAI but I subscribe to Vidu and Typecast, and have given money to Kling, Hailou, and even Gemini in the form of Google Workspace.
So these other guys have products and use cases, which OpenAI has never been able to crack beyond ChatGPT. And ChatGPT was never worth paying for, IMO.
If OpenAI dies, it's not because there is no market for the technology (which is all NVIDIA cares about), it's more that OpenAI doesn't know how to run a relevant technology company.
They were given everything, not just NVIDIA's billions of dollars and credit backing but all the first-mover advantage, all the respect and credibility early on, so it's really sad to see them unable to develop interesting products and turn a profit in a space they helped pioneer, while so many others are making money with the tech all around them.
NVIDIA is fine. The technology will continue to improve and NVIDIA will stay at the center. OpenAI is fucked - knew it when they retired Sora to focus on text-to-text and coding (a largely solved problem).
Do you use coding agents? Just curious bc from my experience using coding agents, open ai’s codex is neck and neck with anthropic’s claude code if not ahead. I wouldn’t agree that OpenAI hasn’t done anything since ChatGPT since codex is my daily driver for software engineering
It's kinda nuts to me how people can act like Claude is lightyears ahead of OpenAI models. Sure, it's one thing to simply have a preference or claim that Claude does some things better, but in reality they are both about as effective at doing the same job. I've long preferred GPT models because they know better how to shut up and don't seem to overthink as much as Claude, but I'm under no illusions that if OpenAI went belly-up then I couldn't do my job essentially the same with Claude. GPT models have also clearly improved over time in terms of programming. There haven't been any "big bangs" necessarily, but it's really not hard to give the same task to 5.3 and 6 and see which one has better output.
1 reply →
To each their own. When OpenAI droped Sora and focused more on Codex, the product improved dramatically and I'm probably not the only one who dumped Claude Code subscription in favor of Codex; OpenAI's is miles ahead of Anthropic and has been since at least the release GPT 5.6 Sol - even the PR and generous resets is far better than how Anthropic is nickel and diming by requiring paid subscribers to pay yet more credits to even use their best available model (which is not even as good as OpenAI's top 2 models)
I've paid more money to OpenAI than I've spent on all other products I pay for combined the past 5 years.
> if OpenAI can’t use the compute, someone else can
How? The hardware is in OpenAI's datacenters. Does Nvidia have a couple hundred semi trucks, contractors, and IT technicians, to repo the hardware and resell it to someone else before it's lost most of its value? These chips will be replaced approx every 3-4 years. So if OpenAI tanks, after Nvidia pays for and waits for the process to collect the hardware, they then have to sell it for pennies on the dollar. They lose almost all the investment.
Also consider that SpaceXAI already had datacenters full of gear that they basically weren't using because nobody wanted their product, so they now rent it to Anthropic. The demand for hardware isn't really there at the scale of OpenAI.
This is the fun part: the AI bubble bursts and the price of components keeps rising. Why? Because companies can just sell you a glorified thin client and force your average user to buy their compute, all subscription-like, from data centers.