Demis Hassabis warns AI hype masks real potential as startups raise tens of billions

Google DeepMind CEO Signals AI Startup Bubble Warning

Demis Hassabis, the Google DeepMind CEO who won a Nobel Prize in Chemistry for his work on AlphaFold, has issued a stark warning about the current state of AI investment. Speaking on a recent episode of Google DeepMind: The Podcast, Hassabis cautioned that parts of the artificial intelligence ecosystem are showing clear signs of a bubble, particularly among early-stage AI companies securing massive funding rounds 1

1

. He pointed to AI startups raising tens of billions in valuations before fully launching their products as a red flag that cannot be ignored 2

2

.

The Alphabet subsidiary leader questioned how these unsustainable valuations could persist when many companies "basically haven't even got going yet" 2

2

. His assessment comes amid an investor frenzy that has propelled the AI startup boom to unprecedented heights, with seed-stage companies commanding valuations that would have been unthinkable just years ago. Hassabis described the phenomenon as "an overreaction to the underreaction," reflecting on how technological shifts often swing rapidly from skepticism to obsession 2

2

.

Big Tech Investment Versus Speculative Funding

Hassabis drew a clear distinction between the speculative funding rounds of early-stage AI companies and the substantial investment pouring in from Big Tech giants like Google, OpenAI, Meta and Anthropic. When it comes to Big Tech spending on artificial intelligence infrastructure and research, he emphasized that "there's a lot of real business underlying that" 1

1

. These established companies are backed by existing revenues, operational infrastructure and long-term strategic roadmaps that make their valuations more credible 3

3

.

The Google DeepMind leader's perspective carries weight given his firsthand experience building one of the field's most respected research organizations. Founded by Hassabis in 2010 and acquired by Alphabet in 2014, DeepMind struggled to attract belief or funding in its early days 3

3

. Now, artificial intelligence dominates global business and policy discussions, creating what Hassabis sees as a market due for a correction as reality catches up to AI hype 3

3

.

Long-Term Potential of AI Remains Underappreciated

Despite his concerns about short-term speculation, Demis Hassabis maintains that AI is "overhyped in the short term and still underappreciated in the medium to long term" 1

1

. His confidence stems from tangible scientific breakthroughs, particularly his Nobel Prize-winning work on AlphaFold, an AI system that predicts protein structures. He also leads Isomorphic Labs, an Alphabet subsidiary focused on applying AI to drug discovery 1

1

.

Google DeepMind is pushing beyond current capabilities toward world models—systems that understand not just language but the physical world itself. Through products like its Genie systems, the lab is developing AI that can grasp how objects move and interact 1

1

. This physical understanding is essential for advances in robotics and universal assistants, representing the kind of foundational work that Hassabis believes will deliver lasting value. The technology could also reshape gaming, a longtime passion of the former video game programmer who once founded Elixir Studios 1

1

.

Intense Competition in AI Race to AGI

Hassabis acknowledged the intense competition in AI as companies race toward Artificial General Intelligence (AGI), describing it as "the most ferocious capitalist competition there's ever been" 1

1

. This contest has unleashed a wave of investment reminiscent of the dot-com boom, with Google competing alongside OpenAI, Meta and Anthropic. Recent developments underscore this pressure: French startup Mistral launched its Mistral 3 AI model suite with open-weight models, while reports emerged that OpenAI CEO Sam Altman declared a "code red" internally due to intensifying competition 2

2

.

Despite the competitive pressure, Hassabis describes himself as friendly "with pretty much all of them," referring to fellow tech executives—a level of harmony that isn't universal in the field 1

1

. Meanwhile, Goldman Sachs Asset Management maintained that the broader AI sector remained healthy despite recent stock declines in companies like CoreWeave. Sung Cho, co-head of public tech investing, noted that 90% of AI infrastructure funding came from corporate cash flows rather than debt, describing recent volatility as isolated incidents rather than systemic risks 2

2

. Hassabis remains focused on developing foundational AI systems at Google DeepMind, including models that power Gemini and the company's frontier AI research 3

3

.