There's a lot of money sloshing around in artificial intelligence right now, and not everyone thinks that's going to end well. Demis Hassabis, who runs Alphabet Inc.'s (GOOGL) (GOOG) Google DeepMind division, is raising a yellow flag about AI startup valuations that seem detached from reality.

The Pre-Launch Funding Phenomenon

Speaking on "Google DeepMind: The Podcast" this week, Hassabis pointed out something that should make investors pause: AI startups are securing tens of billions in valuations before they've actually, you know, launched anything substantial.

"Some startups basically haven't even got going yet," he said. "It's sort of interesting to see how can that be sustainable. You know, my guess is probably not, at least not in general."

It's a classic pattern in technology. When something genuinely transformative emerges, the pendulum swings hard from skepticism to euphoria, and valuations inflate faster than you can say "disruptive innovation."

Tech Giants vs. Startup Hype

Hassabis drew a clear line between these hyped seed-stage ventures and companies like Google that have "a lot of real business" supporting their AI ambitions. When you're sitting on massive revenue streams from existing products, your AI investments look a lot less risky than when you're burning through venture capital without a clear path to profitability.

He offered a nuanced take on AI's trajectory: it's "overhyped in the short term" but "still underappreciated in the medium to long-term." That's the kind of statement that makes everyone think they're smart enough to time the market perfectly, which is usually how bubbles get bigger.

The DeepMind chief noted that these swings are almost inevitable. "It's almost an overreaction to the underreaction," he said, reflecting on his company's early days when convincing people to take AI seriously was an uphill battle. Now the problem is the opposite.

Is There Actually a Bubble?

Not everyone agrees we're heading for a crash. Despite recent stock declines at companies like CoreWeave Inc. (CRWV), Goldman Sachs Asset Management argues the broader AI sector is on solid ground.

Sung Cho, co-head of public tech investing at Goldman, pushed back against concerns about a debt-fueled bubble. The key detail: 90% of AI infrastructure funding comes from corporate cash flows, not borrowed money. That's a fundamentally different dynamic than classic bubble scenarios. Recent stock drops and supply chain hiccups at some firms, he said, are isolated problems rather than signs of systemic rot.

Meanwhile, Competition Heats Up

While Hassabis was cautioning about valuations, French startup Mistral was busy making waves with its Mistral 3 AI model suite. The launch includes a large multimodal, multilingual model plus nine smaller customizable versions.

Mistral's approach is notable: by releasing open-weight models, they're letting users run AI on their own hardware, positioning themselves as an alternative to OpenAI and Google's more proprietary approaches.

The timing is interesting. Reports suggest that OpenAI CEO Sam Altman recently declared an internal "code red" due to intensifying competition. When you're the market leader and you're declaring emergencies, that tells you something about how fast this space is moving.

Whether we're in a bubble or not, one thing is clear: the AI market is moving at a pace that makes traditional valuation methods look quaint. Hassabis might be right that some corrections are coming, but predicting when and how severe is another matter entirely.