Anthropic CEO weighs in on AI bubble talk and risk-taking among competitors

Anthropic CEO Dario Amodei shared his thoughts on whether the AI ​​industry was in a bubble in The New York Times DealBook Summit on Wednesday. This was in addition to throwing shade at a certain unnamed competitor, which was clearly OpenAI.

Amodei refused to give a simple yes-or-no answer to the question about a bubble, saying it was a complex situation, but instead explained his thoughts on the economics of AI in more detail.

He described himself as optimistic about the technology’s potential, but warned that there could be players in the ecosystem who could make a “timing error” or see “bad things” happen when it comes to economic outcomes.

“There is an inherent risk when the timing of economic value is uncertain,” Amodei explains. He said companies had to take risks to compete with each other and with authoritarian opponents — a reference to the threat from China — but added that some players were “not managing that risk well and are taking unwise risks.”

The problem, he said, is the uncertainty about how quickly the economic value of AI will grow and that properly identifying the lag times in building more data centers.

‘There is [a] A real dilemma, which we as a company are trying to deal with as responsibly as possible,” said Amodei. “And then I think there are some players who are ‘YOLOing’, who are pushing the risk dial too far, and I am very concerned,” he added, using the slang term for “you only live once,” which is often used to justify taking risks.

Additionally, he addressed the question surrounding the timelines for the retirement of AI chips. That’s another hot topic and a factor that could negatively impact the economics of the industry if GPUs become obsolete and lose their value earlier than planned.

“The problem is not the lifespan of the chips; chips continue to work for a long time. The problem is that new chips are coming onto the market that are faster and cheaper… and so the value of old chips may drop somewhat,” Amodei said.

He said Anthopic was making conservative assumptions on this and other fronts as it faces an uncertain future.

The AI ​​company’s revenue has grown 10 times a year for the past three years, the CEO said, from zero to $100 million in 2023, then from $100 million to $1 billion in 2024, and will reach somewhere between $8 billion and 10 billion by the end of this year.

But Amodei said he would be “really stupid” if he assumed this pattern would continue. “I don’t know if it’s going to be 20 billion in a year or 50… it’s very uncertain. I’m trying to plan conservatively. So I’m planning the lower end of it, but that’s very worrying,” he said.

AI companies like his need to plan how much computing power they will need in the coming years, and how much they will need to invest in data centers. If they don’t buy enough, they may not be able to serve their customers. And if they buy too much, they will struggle to keep up with costs, or in the worst case, they could go bankrupt.

Last month, OpenAI landed in a PR crisis when its CFO said she wanted the U.S. government to “backstop,” or insure, her company’s infrastructure loans so that taxpayers would pick up the tab if OpenAI couldn’t. After the furor, she walked back the comments.

Those who take more risks could be overextending themselves, Amodei warned, especially if “you’re someone who, like the Constitution, just wants to ‘YOLO’ things, or just likes big numbers,” he said in a veiled reference to OpenAI CEO Sam Altman.

“We think that in principle everything is going well in almost all worlds… I can’t speak for other companies,” he said.