Dario Amodei, CEO of Anthropic, shared his thoughts on whether the AI industry is in a bubble within the New York Times DealBook Summit on Wednesday. This further solid a shadow over a certain unnamed competitor, which was clearly OpenAI.
Amodei declined to provide an easy yes-or-no answer to the query a few bubble, saying it’s a posh situation, but as a substitute explained his thoughts on the economics of AI in additional detail.
He described himself as optimistic concerning the technology's potential, but warned that there might be players within the ecosystem who could make a “mistake in timing” or see “bad things” on the subject of the economic advantages.
“There is an inherent risk when the timing of the economic value is uncertain,” Amodei explained. He said corporations have to take risks to compete with one another and with authoritarian opponents – a reference to the threat posed by China – but added that some players are “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 appropriately attributing those aspects to the delays in constructing more data centers.
“There is an actual dilemma that we as an organization try to handle as responsibly as possible,” Amodei said. “And then I feel there are some players who’re 'YOLOing,' who’re pushing the danger button too far, and I'm very concerned,” he added, using the slang term for “you simply live once,” which is commonly used to justify taking risks.
He also addressed the query of the discontinuation periods for AI chips. This is one other hot topic and an element that might negatively impact the industry's economics if GPUs change into obsolete and lose value prematurely.
Techcrunch event
San Francisco
|
thirteenth–fifteenth October 2026
“The problem will not be the lifespan of the chips – chips work for a very long time. The problem is that latest chips are coming onto the market which are faster and cheaper… and due to this fact the worth of old chips may decrease somewhat,” Amodei said.
He said Anthropic made conservative assumptions on this and other fronts as the corporate planned for an uncertain future.
The AI company's revenue has grown 10-fold per yr over the past three years, the CEO said, from zero to $100 million in 2023, then from $100 million to $1 billion in 2024, and shall be somewhere between $8 billion and $10 billion by the tip of this yr.
But Amodei said it could be “really silly” to easily assume the pattern would proceed. “I don't know if it can be 20 billion or 50 billion in a yr… it's very uncertain. I'm attempting to plan conservatively. So I'm planning for the lower sure, but that's very worrying,” he said.
AI corporations like his have to plan how much computing power they may need in the approaching years and the way much they need to put money into data centers. If they don't buy enough, they could not have the option to serve their customers. And in the event that they buy an excessive amount of, they may struggle to maintain up with costs, or within the worst case scenario, they may go bankrupt.
Last month, OpenAI found itself in a public relations crisis when its CFO said it wanted the U.S. government to “secure” its company's infrastructure loans in order that taxpayers would foot the bill if OpenAI couldn't. After the furor, she walked back the comments.
Those who take more risks could overextend themselves, Amodei warned, especially if “you're a one who just, like constitutionally, just wants 'YOLO' things or simply likes big numbers,” he said, in a veiled reference to OpenAI CEO Sam Altman.
“We think we will principally get by in almost any world… I can't speak for other corporations,” he said.
