Nvidia pitches insurers to backstop neocloud GPU loans

Nvidia has moved from a niche processor maker to one of the world's most valuable companies on the back of an AI arms race that drove nearly insatiable demand for its chips. In the early days of the boom, almost all of that demand came from investment-grade tech giants like Amazon, Microsoft, Google and Meta that could fund chip purchases from their own balance sheets. Over the past 24 months, a growing share of demand has come instead from neoclouds, young companies with short track records that almost always need to borrow to buy GPUs and build data centers, and lenders have been wary of that risk profile.

The proposed insurance structure is a response to that wariness. Neocloud financing already leans on chips and data centers as collateral, but chips are regarded as short-lived assets that depreciate relatively quickly, and skeptics of Big Tech's AI spending often cite that depreciation as a flaw in the boom's economic assumptions. Nvidia CEO Jensen Huang has argued chips have long useful lives and should be regarded as an "investable asset class." Nvidia has reportedly supplied at least one insurer with information on chip depreciation and future computing capacity valuation, and is working with insurance broker Howden Re. Because the potential transactions would be enormous, beyond individual insurers' balance sheets, Nvidia is reportedly exploring syndicating risk to hedge funds and other alternative investors, and has considered participating itself.

This fits a broader pattern of Nvidia using its financial might to reinforce demand for its products. The $35B Lambda–Anthropic cloud deal, backstopped by Nvidia, is another example. Such deals have raised red flags among some investors, drawing accusations of circular financing and comparisons to dot-com era vendor financing, though data center industry leaders generally dispute that characterization. The stakes may hinge on whether insurers and alternative investors are willing to price depreciation risk on assets whose resale value is hard to predict, and on how much of the risk ends up linked to hyperscalers, institutional investors and debt markets.