The AI gold rush has reached a peculiar stage where startups are securing tens of billions in valuations armed with little more than pitch decks and ambition. Even one of artificial intelligence's pioneering figures thinks something doesn't add up.
Demis Hassabis, co-founder and CEO of Google DeepMind, recently voiced skepticism about how money is flowing into AI companies that haven't proven much of anything yet. Speaking on "Google DeepMind: The Podcast," he pointed to startups raising at sky-high valuations despite having limited traction.
"They're raising at tens of billions of dollars in valuations just out of the gate," Hassabis said. "It's sort of interesting to see, can that be sustainable? My guess is, probably not, at least not in general."
When the Pitch Deck Is Worth More Than the Product
Hassabis told podcast host Hannah Fry that some seed-stage startups "basically haven't even got going yet" while securing funding rounds that would make a mature company blush. He drew a clear line between these frothy early-stage deals and the capital flowing into established tech giants building AI infrastructure.
At big tech firms, he said, there's "a lot of real business" supporting those valuations. These companies have products, users, and revenue streams that anchor their AI investments to reality. The seed-stage startups raising billions? Not so much.
Hassabis also offered his now-familiar view that AI is "overhyped in the short term" while "still underappreciated in the medium to long term." He described this imbalance as typical of how transformative technologies develop.
From Skepticism to Euphoria in One Decade
The sentiment shift around AI has been dramatic. "When we started DeepMind, no one believed in it," Hassabis recalled. "Fast forward 10, 15 years, and now, obviously, it seems to be the only thing people talk about in business."
That rapid reversal can create strange market dynamics. "It's almost an overreaction to the underreaction," Hassabis said, suggesting that an "overcorrection" often follows when technologies move quickly from widespread skepticism to intense focus. The result? Valuations that leap ahead of fundamentals.
Focus on the Long Game
Despite his observations about market froth, Hassabis said he's not particularly interested in debating whether AI is in a bubble. His attention remains on long-term development at Google DeepMind rather than short-term valuation noise.
He told Fry that AI investment by large technology companies makes sense because it's tied to existing products, infrastructure, and ongoing research programs. There's a foundation underneath all that spending.
Other Voices Raising Caution Flags
Hassabis isn't alone in questioning AI valuations. Howard Marks, co-founder of Oaktree Capital Management, raised similar concerns in a recent memo, pointing to uncertainty around pricing and long-term returns in AI investments.
Meanwhile, industry leaders are trying to calibrate expectations about AI's impact on work. "Everybody's jobs will be different," Nvidia (NVDA) CEO Jensen Huang said during a panel discussion at the U.S.–Saudi Investment Forum in Washington, D.C., in November. He suggested AI tools will augment everyday workflows rather than fully replace workers.
The question hanging over all this enthusiasm and capital is straightforward: Can companies with no revenue and minimal traction really be worth billions? Hassabis has offered his answer. Investors will eventually provide theirs.