Anthropic’s CEO says AI boom is real, but some companies are YOLO-ing

Just a heads up, if you buy something through our links, we may get a small share of the sale. It’s one of the ways we keep the lights on here. Click here for more.

At The New York Times DealBook Summit this week, Anthropic CEO Dario Amodei showed up ready to talk economics, geopolitics, and apparently,  subtweet the entire OpenAI executive team without ever saying their name. 

When asked if the AI industry is in a bubble, Amodei avoided a simple yes/no answer, opting instead for a vibe-check analysis: the future is promising, but also potentially chaotic, depending on how many CEOs decide to spend billions like they’re playing tech-industry roulette.

Amodei said he’s optimistic about AI’s long-term potential, but warned that some companies are making what he politely called “timing errors” and less politely referred to as “YOLO-ing.” 

In AI-speak, that means building massive data centers before anyone is totally sure the revenue curves will match the electric bills. 

He also nodded to global competition, particularly with China, noting that companies feel pressure to move fast, sometimes faster than good judgment recommends.

Another looming risk: chip depreciation. GPUs don’t actually die. They just become embarrassingly slow compared to the shiny new models. 

That means billions spent today might be worth half as much tomorrow.

Anthropic, he said, is taking a more conservative approach rather than panic-buying hardware like it’s Black Friday and the apocalypse at the same time.

Still, business is booming, for now. 

Anthropic’s revenue has reportedly grown 10x annually for three years straight, going from zero to $100 million, then $1 billion, and now maybe $8-10 billion by year’s end. 

But Amodei insists he’s not assuming infinite exponential growth. “I’d be really dumb” to count on that, he admitted.

Planning infrastructure in this environment, he explained, feels like choosing between running out of servers or accidentally bankrupting yourself. 

And while he didn’t name names, he took a clear jab at OpenAI after its CFO floated the idea of government-insured loans, a suggestion that did not go over great.

Anthropic’s strategy? Grow, but maybe don’t set the money printer to “YOLO.” 

As Amodei put it: “We think we’re going to be okay. I can’t speak for other companies.”