Jensen Huang Says AI Cloud Demand Is Booming—Markets See a Debt Trap
Jensen Huang Says AI Cloud Demand Is Booming—Markets See a Debt Trap
At Nvidia’s GTC gathering in Berlin, Huang’s message was relentless: build faster. On Wednesday, he unveiled a containerized system designed to bring AI computing online more quickly, arguing that delays can cost billions. He put the price of a one-gigawatt data center at roughly $50 billion—a reminder that the race to supply AI is also a race to fund it.
The spectacle also had a lighter side. Huang spent part of his keynote directing tiny Microduck robots, a demonstration that prompted Hugging Face chief executive Clement Delangue to hail the ducks as a European creation backed by “American capital and ambition.” But behind the showmanship, the mood around neoclouds—the specialist providers that lease AI computing power—was darker.
By Thursday, Huang told CNBC that demand was “off the charts” and that neocloud capabilities were “extraordinary.” His diagnosis was not a lack of customers but a funding squeeze: “What is missing in the middle is the working capital to fuel that growth.” He also singled out Europe’s expensive, scarce power as a constraint, arguing that local infrastructure is essential if the continent is to keep the resulting AI economy at home.
Markets are reading the same facts far less generously. CoreWeave and Nebius shares fell as investors questioned whether debt-heavy data-center construction can generate returns quickly enough. CoreWeave had $35.6 billion in debt as of June, while negative analyst calls and rising borrowing costs intensified scrutiny of the model. Huang, who did not name the company in his Thursday remarks, maintained that capital scarcity would be solved and said Nvidia hardware was becoming a financeable asset class.
That leaves a sharp divide: Nvidia sees a temporary bottleneck in a durable boom; skeptical investors see the bill arriving before the payoff.
Write a comment