Demis Hassabis, co-founder and CEO of DeepMind, has issued a cautionary note about the current state of AI startup valuations, warning that some are becoming unsustainable. Speaking in an episode of "Google DeepMind: The Podcast" released Tuesday, Hassabis expressed concern that a bubble may be forming in the early-stage AI startup ecosystem. He pointed out that many startups are securing massive funding at valuations in the tens of billions of dollars—often before they’ve even launched a product or demonstrated meaningful progress. “Some startups basically haven’t even got going yet,” Hassabis said, “yet they’re raising at these enormous valuations just out of the gate.” He questioned how such valuations could be sustained, adding, “My guess is probably not, at least not in general.” Hassabis made a clear distinction between these early-stage ventures and the major tech companies investing heavily in AI infrastructure. While he acknowledged that Big Tech’s AI spending is backed by real business operations and existing revenue streams, he warned that the current enthusiasm for AI startups may be driven more by hype than substance. He described the rapid shift from widespread skepticism to near-obsession with AI as a classic pattern in technological revolutions. “When we started DeepMind, no one believed in it,” he recalled. “Fast forward 10, 15 years, and now, obviously, it seems to be the only thing people talk about in business.” He noted that such swings often lead to overvaluation, calling it “an overreaction to the underreaction.” Despite his concerns about market dynamics, Hassabis said he isn’t personally worried about whether AI is in a bubble. His focus remains on DeepMind’s mission: developing cutting-edge AI models that power Google’s products like Gemini and advancing frontier research. The warning comes amid a surge in AI startup funding. Business Insider recently reported that young founders—some still in college or fresh out—have raised millions with minimal track records. One Stanford dropout, Carina Hong, raised $64 million for her AI math startup, Axiom Math, recruiting top talent from Meta and Google Brain. Across 16 young founders interviewed by Business Insider, over $100 million in funding has been secured this year. Still, not all investors are convinced. Howard Marks, co-founder of Oaktree Capital Management, questioned the logic of betting on startups with no revenue or profits. “Do you want to invest in a novel entrepreneurial startup with no revenues and no profits today, but could be a moonshot if it works?” he asked on the “We Study Billionaires” podcast. “Or do you want to invest in a great tech company that’s already making money, where AI is just an incremental boost?” The choice, he said, is a fundamental one.