The CEO of generative AI firm Anthropic warned that the purchase of escalating data center compute costs could be "ruinous" for the company if timed wrong.

Speaking on the Dwarkesh Podcast, Dario Amodei appeared to contrast the company's data center spend to the more profligate OpenAI – although the rival was never mentioned by name.

Asked why, if he truly believed data centers would soon be equivalent to Nobel Prize Winners, Anthropic was not spending even more on its data centers, Amodei said that those systems would take time to turn a profit.

"I really do believe that we could have models that are a country of geniuses in the data center in one to two years," he said.

"One question is: How many years after that do the trillions in revenue start rolling in? I don’t think it’s guaranteed that it’s going to be immediate. It could be one year, it could be two years, I could even stretch it to five years, although I’m sceptical of that."

This uncertainty means that "with the way you buy these data centers, if you’re off by a couple years, that can be ruinous," he said.

Using pharmaceutical breakthroughs as a potential outcome of a genius AI, Amodei said that it would take years for any wonder-drugs to be developed, get regulatory approval, and be commercialized – and that's with the revenue split between Anthropic and the pharmaceutical company unclear.

However, in this case, if Anthropic spent heavily on compute too early, it may not survive long enough to benefit from the drug's rollout.

"So when we go to buying data centers, again, the curve I’m looking at is: we’ve had a 10x a year increase every year," he said. "At the beginning of this year, we’re looking at $10 billion in annualized revenue. We have to decide how much compute to buy. It takes a year or two to actually build out the data centers, to reserve the data center.

"Basically, I’m saying, 'In 2027, how much compute do I get?' I could assume that the revenue will continue growing 10x a year, so it’ll be $100 billion at the end of 2026 and $1 trillion at the end of 2027. Actually, it would be $5 trillion dollars of compute because it would be $1 trillion a year for five years."

This means that the company "could buy $1 trillion of compute that starts at the end of 2027. [But] if my revenue is not $1 trillion dollars, if it’s even $800 billion, there’s no force on earth, there’s no hedge on earth that could stop me from going bankrupt if I buy that much compute."

Amodei admitted that "part of my brain wonders if it’s going to keep growing 10x," but if he is "just off by a year in that rate of growth, or if the growth rate is 5x a year instead of 10x a year, then you go bankrupt."

"You end up in a world where you’re supporting hundreds of billions, not trillions. You accept some risk that there’s so much demand that you can’t support the revenue, and you accept some risk that you got it wrong, and it’s still slow."

Last November, Sam Altman, CEO of rival OpenAI, said that his company had more than $1.4 trillion in data center commitments to support its AI buildout. OpenAI has since announced further compute deals.

"I think it is true we’re spending somewhat less than some of the other players," Amodei said. "[Being responsible means] have we been thoughtful about it or are we YOLOing and saying, 'We’re going to do $100 billion here or $100 billion there?' I get the impression that some of the other companies have not written down the spreadsheet, that they don’t really understand the risks they’re taking. They’re just doing stuff because it sounds cool."

Asked about the risk of being beaten by higher spending rivals, he countered that "there’s this chance that they go bankrupt before. Again, if you’re off by only a year, you destroy yourselves. That’s the balance. We’re buying a lot. We’re buying a hell of a lot. We’re buying an amount that’s comparable to what the biggest players in the game are buying."

Anthropic has long used Google Cloud for some of its model work, and last October announced a 1GW+ cloud deal with up to 1 million Google TPUs.

The generative AI company separately announced a $50bn data center investment at Fluidstack facilities – although they are the same sites Google operated out of, so it may be a related deal.

Anthropic investor Amazon launched the Project Rainier cluster featuring its own Trainium2 chips for use by the AI firm last year. It plans to spend as much as $11 billion on the site at full buildout.

In November, the company announced it would spend $30bn on Microsoft Azure credits - alongside a $5bn Microsoft investment.

Amodei continued: "But if you’re asking me, 'Why haven’t we signed $10 trillion of compute starting in mid-2027?'... First of all, it can’t be produced. There isn’t that much in the world.

"But second, what if the country of geniuses comes, but it comes in mid-2028 instead of mid-2027? You go bankrupt."