Nvidia is pitching a new financing structure for the companies racing to build AI datacenters, according to The Register, which describes the arrangement as a "double-dipping" scheme.
The report points to the money problem behind the AI boom. To keep pace with demand for computing power, so-called rent-a-GPU firms — companies that buy Nvidia chips and rent out access to them — have had to raise enormous sums. According to The Register, outfits such as CoreWeave and Lambda have borrowed billions of dollars from venture capitalists and hedge funds to bankroll their datacenter buildouts.
That borrowing works only as long as the underlying business keeps performing, which is why how these buildouts get financed has become a closely watched question. The Register frames Nvidia's proposed structure as a way to help keep that capital flowing.
The available reporting is a headline and short summary rather than a full breakdown of the mechanics, so the precise terms of Nvidia's proposal aren't spelled out in the source material beyond the "double-dipping" characterization.
Why it matters: Nvidia both sells the chips and now appears to be shaping how buyers pay for the datacenters that house them — a sign of how deeply one company sits at the center of the AI infrastructure economy, and of the growing financial engineering required to sustain the buildout.