Demis Hassabis warns that AI startup valuations could be headed for a correction
Google DeepMind’s cofounder and CEO, Demis Hassabis, isn’t shy about calling out what he sees as inflated pricing in the AI startup scene. In a recent episode of “Google DeepMind: The Podcast,” he described a bubble forming around early-stage AI ventures that are being valued at tens of billions of dollars before they have even proven their business models.
Some startups, he noted, haven’t “even got going yet,” yet they’re securing funding at sky-high valuations. He questioned how such a trajectory could be sustainable, suggesting that, in general, it probably isn’t.
Hassabis drew a line between these exuberant seed rounds and the substantial investments flowing into AI infrastructure by large tech companies. He argued there is a solid, real business foundation behind Big Tech’s AI-related valuations, whereas the younger, flashier startups may not have comparable fortitude.
While acknowledging that AI is overhyped in the near term, he maintained that the technology remains undervalued in the medium and long term. He warned that a broad over-correction is a common feature of major tech shifts, particularly when public sentiment swings rapidly from skepticism to obsession.
Reflecting on his own journey, Hassabis recalled that when DeepMind started, few believed in the venture. Over the next decade or so, it became widely recognized as a central topic in business strategy. He described this dramatic swing as often pushing valuations beyond what fundamentals would justify, an overreaction to prior underappreciation.
Despite concerns about a possible bubble, Hassabis said he isn’t preoccupied with market hype. His focus remains on DeepMind’s core mission: developing AI models that power Google’s products, including the Gemini project, and advancing frontier AI research.
Rising valuations in AI startups continue to draw attention
Recent reporting shows AI startups drawing increasingly high valuations, with many young founders—some recently out of school—raising substantial sums. Stories include a Stanford dropout who raised $64 million for an AI-enabled math startup, Axiom Math, and the recruitment of top-tier AI talent from industry giants.
In total, a cohort of about 16 young founders talked to this publication has secured more than $100 million in funding this year. Yet not everyone is buying into the hype. Howard Marks, cofounder of Oaktree Capital Management, warned that investors are piling into AI ventures with little track record, prompting questions about whether such bets are prudent.
He framed the choice this way: invest in a novel, potentially moonshot startup with no revenues today, or back a proven tech company that already generates substantial profits where AI could yield incremental gains rather than transformative shifts. The conversation invites readers to weigh which path they believe offers the best risk-adjusted opportunities in the evolving AI landscape.
Would you agree that AI hype is masking underlying risk, or do you see genuine, sustainable growth driving these high valuations? Share your perspective in the comments.
