NVIDIA might be better positioned to first get to/first scale up access to AGIs than any of the AI labs that typically come to mind. They’re already the world’s highest-market-cap company, have huge and increasing quarterly income (and profit) streams, and can get access to the world’s best AI hardware at literally the best price (the production cost they pay). Given that access to hardware seems far more constraining of an input than e.g. algorithms or data, when AI becomes much more valuable because it can replace larger portions of human workers, they should be highly motivated to use large numbers of GPUs themselves and train their own AGIs, rather than e.g. sell their GPUs and buy AGI access from competitors. Especially since poaching talented AGI researchers would probably (still) be much cheaper than building up the hardware required for the training runs (e.g. see Meta’s recent hiring spree); and since access to compute is already an important factor in algorithmic progress and AIs will likely increasingly be able to substitute top human researchers for algorithmic progress. Similarly, since the AI software is a complementary good to the hardware they sell, they should be highly motivated to be able to produce their own in-house, and sell it as a package with their hardware (rather than have to rely on AGI labs to build the software that makes the hardware useful).
This possibility seems to me wildly underconsidered/underdiscussed, at least in public.
When Nvidia did make their own nemotron models, they performed much worse than other models of the same size, despite using a lot of compute. I think this type of research is not their strength.
I am skeptical of this because they can’t just scale up data centers on a dime. And signaling that they are trying to become the new biggest hyperscaler would be risky for their existing sales: big tech and frontier labs will go even harder for custom chips than they are now.
To make this happen Nvidia would probably need to partner with neoclouds like CoreWeave that have weaker affiliations with frontier labs. Nvidia is actively incubating neoclouds and does have very strong relationships here, to be sure, but the neoclouds still have fewer data centers and less technical expertise than the more established hyperscalers.
And I think algorithms and talent are very important.
And signaling that they are trying to become the biggest new hyperscaler would be risky for their existing sales: big tech and frontier labs will go even harder for custom chips than they are now.
In the world that I see Bogdan as pointing to, ‘when AI becomes much more valuable because it can replace larger portions of human workers’, I’d expect revenue from supplying AGIs to strongly outweigh revenue from chip sales.
I’m saying it would be challenging for Nvidia to preserve its high share of AI compute production in the first place while trying to execute this strategy. Nvidia is fabless, and its dominance will erode if labs/hyperscalers/Broadcom create satisfactory designs and are willing to place sufficient large orders with TSMC.
Nvidia already has an AI cloud division that is not negligible but small compared to the big players. But they appear to not even own their own chips: they lease from Oracle.
Conventional wisdom suggests “execution” for hyperscale consumer products is a moat, e.g “Apple may lead scaling access to AGIs since they have the design, supply chain, marketing expertise plus a vast, established user ecosystem (>2bn active devices)”. AGI, however, dissolves away the edge from expertise, and users will flock to a new thing if the value is there (ChatGPT surpassed 1m users in 5 days).
A counter idea I have though is that a prerequisite for AGI may be access to training data derived from pre-AGI systems being used in the wild (e.g across 2bn active devices). In this case, NVIDIA might not have access to the data required to come first.
This is the kind of post that would benefit from an “epistemic status.”
If no one thinks NVIDIA is competitive at building frontier systems there’s probably a reason. How seriously should I take these informal arguments without citations? Are they just speculation?
NVIDIA might be better positioned to first get to/first scale up access to AGIs than any of the AI labs that typically come to mind.
They’re already the world’s highest-market-cap company, have huge and increasing quarterly income (and profit) streams, and can get access to the world’s best AI hardware at literally the best price (the production cost they pay). Given that access to hardware seems far more constraining of an input than e.g. algorithms or data, when AI becomes much more valuable because it can replace larger portions of human workers, they should be highly motivated to use large numbers of GPUs themselves and train their own AGIs, rather than e.g. sell their GPUs and buy AGI access from competitors. Especially since poaching talented AGI researchers would probably (still) be much cheaper than building up the hardware required for the training runs (e.g. see Meta’s recent hiring spree); and since access to compute is already an important factor in algorithmic progress and AIs will likely increasingly be able to substitute top human researchers for algorithmic progress. Similarly, since the AI software is a complementary good to the hardware they sell, they should be highly motivated to be able to produce their own in-house, and sell it as a package with their hardware (rather than have to rely on AGI labs to build the software that makes the hardware useful).
This possibility seems to me wildly underconsidered/underdiscussed, at least in public.
When Nvidia did make their own nemotron models, they performed much worse than other models of the same size, despite using a lot of compute. I think this type of research is not their strength.
I am skeptical of this because they can’t just scale up data centers on a dime. And signaling that they are trying to become the new biggest hyperscaler would be risky for their existing sales: big tech and frontier labs will go even harder for custom chips than they are now.
To make this happen Nvidia would probably need to partner with neoclouds like CoreWeave that have weaker affiliations with frontier labs. Nvidia is actively incubating neoclouds and does have very strong relationships here, to be sure, but the neoclouds still have fewer data centers and less technical expertise than the more established hyperscalers.
And I think algorithms and talent are very important.
In the world that I see Bogdan as pointing to, ‘when AI becomes much more valuable because it can replace larger portions of human workers’, I’d expect revenue from supplying AGIs to strongly outweigh revenue from chip sales.EDIT: I misunderstood Josh’s point.
I’m saying it would be challenging for Nvidia to preserve its high share of AI compute production in the first place while trying to execute this strategy. Nvidia is fabless, and its dominance will erode if labs/hyperscalers/Broadcom create satisfactory designs and are willing to place sufficient large orders with TSMC.
Got it! I misunderstood you.
Nvidia already has an AI cloud division that is not negligible but small compared to the big players. But they appear to not even own their own chips: they lease from Oracle.
Yeah.
Conventional wisdom suggests “execution” for hyperscale consumer products is a moat, e.g “Apple may lead scaling access to AGIs since they have the design, supply chain, marketing expertise plus a vast, established user ecosystem (>2bn active devices)”. AGI, however, dissolves away the edge from expertise, and users will flock to a new thing if the value is there (ChatGPT surpassed 1m users in 5 days).
A counter idea I have though is that a prerequisite for AGI may be access to training data derived from pre-AGI systems being used in the wild (e.g across 2bn active devices). In this case, NVIDIA might not have access to the data required to come first.
This is the kind of post that would benefit from an “epistemic status.”
If no one thinks NVIDIA is competitive at building frontier systems there’s probably a reason. How seriously should I take these informal arguments without citations? Are they just speculation?