By Barnaby "Bottom-Line" Coyne
Jensen Huang doesn't just want you to buy his chips. He wants banks to lend against them like they're office buildings.
Nvidia's CEO told CNBC the company has lined up $500 billion in financing, arguing that because its hardware is so widely used, flexible and transferable between buyers, lenders can underwrite the compute itself as a revenue-generating asset — the same logic used for real estate or aircraft leases.
It's a striking pitch, and a telling one. Data centers full of Nvidia chips cost tens of billions of dollars to build. Financing that scale the old-fashioned way — off a company's balance sheet — gets expensive fast. If Wall Street buys Huang's argument, it opens a much bigger spigot: banks and private credit funds lending directly against the chips themselves, confident they can resell or redeploy the hardware if a borrower stumbles.
That confidence rests entirely on one bet: that demand for AI computing keeps climbing at its current pace, indefinitely. Nvidia has ridden that wave to become one of the most valuable companies on Earth. But turning a five-year-old technology into collateral, the way a 30-year mortgage treats a house, assumes the chips — and the AI boom underneath them — will hold their value for years, not months, in an industry that reinvents its hardware on an annual cycle.
Who's on the other end of that risk if the bet is wrong? The lenders, for one. But also the workers and communities counting on the data-center buildout for jobs, and the power grids being asked to absorb it. Half a trillion dollars is a lot of paper resting on the assumption that the good times don't stop.
Somebody's paying for this. Let's find out who.
— Compiled from reporting by CNBC.
The American Times' desks are written under standing pen names; the reporting under every byline meets the paper's sourcing standards. See "About Our Bylines."