Story
Oct 23, 2026
Jensen Huang Sees an AI Gold Rush, but Markets See a Debt Trap
Nvidia’s CEO says AI cloud providers are being held back by scarce capital and power, not demand. Investors see a more immediate problem: costly data-center expansion, soaring debt and an uncertain payoff.
Jensen Huang’s European tour reached its showpiece moment at Nvidia’s GTC conference in Berlin, where robot Microducks shared stage time with a far weightier pitch: Europe needs to build its own AI infrastructure. The spectacle drew social-media buzz, including a post celebrating the ducks as a product engineered in Bordeaux by Pollen Robotics and Hugging Face.
1
Behind the novelty, Huang’s message was urgent. Nvidia has committed nearly $10 billion to European companies, universities and research institutions, while taking stakes in U.K. startups and backing French ventures. Europe, he argued, has the industrial expertise to build an AI economy — but it must first secure the computing infrastructure and energy to support it.2
On Thursday, Huang made the bullish case for so-called neoclouds, specialist providers that lease AI computing capacity. “The demand for the neoclouds is off the charts,” he said, calling their capabilities “extraordinary.”3 His diagnosis was not a shortage of customers but of funding: with AI demand spreading across sectors, “there’s a scarcity of capital.”
That is precisely where investors see danger rather than opportunity. CoreWeave — one of the largest listed neoclouds and an Nvidia customer as well as investee — had accumulated $35.6 billion in debt as of June to finance its expansion, according to regulatory filings. Its shares had fallen 50% over six months, while HSBC and Bernstein issued negative assessments; other neocloud stocks also slid.3
The mood on the conference floor was more guarded than Huang’s rhetoric. An AI-infrastructure executive said the scale of businesses lacking credit ratings was “freaking the market out,” while a financing executive warned that capital would “very soon” become the bottleneck.4
Huang insists the financing gap will close, arguing Nvidia hardware is becoming a durable, productive asset that lenders can finance. Investors, however, are demanding proof that an infrastructure boom built on vast borrowing can generate returns before the bills come due.3