Nvidia / 29 September 2026

Nvidia Explores Insurance for GPU Backed Lending

Nvidia has approached insurers about sharing the risk of loans secured by its AI chips, according to the Financial Times. One structure under discussion would protect lenders to smaller cloud providers, often called neoclouds, if a borrower defaulted and the pledged GPUs fetched too little on resale to repay the debt. The talks are being led by Ingemar Lanevi, Nvidia’s head of financial solutions, and include work with reinsurance broker Howden Re, the report says. The prospective arrangement addresses a financing constraint for customers whose orders can be large relative to their own balance sheets. Nvidia has not announced an insurance contract; the substantive move is its attempt to bring insurers and other risk-bearing investors into the market for GPU-backed infrastructure finance.

The value of a GPU cluster after several years of use sits at the heart of such lending. A lender can assess current customer contracts and expected computing income, yet it also needs to estimate what the chips would recover if the operator fails. The Financial Times reports that Nvidia has shared chip depreciation and future computing-price data with at least one insurer. Nvidia described AI infrastructure as an investable asset class because it is “uniquely productive, durable and fungible.” Those properties have to be priced in a market where new processor generations arrive frequently and the use of older systems changes over time. More transparent valuation data could let lenders and insurers quote terms for GPU assets with greater confidence, expanding the funding options available to cloud operators.

The report says Nvidia has explored structures in which insurance groups pass part of the exposure to hedge funds or other alternative investors; Nvidia itself has also considered joining a consortium behind agreements. That would broaden the capital base beyond conventional bank lenders and the balance sheets of individual insurers. The products resemble residual-value insurance, which pays when equipment is worth less than a specified amount. Specialists such as Forward Compute are already developing that market. Its chief executive, Quentin Saleur, told the Financial Times that insurance could help smaller cloud providers compete for large buyers by reducing concern about their ability to fulfill contracts. For a cloud operator seeking to buy or lease Nvidia systems, the cost and availability of this protection could become part of the economics of its next cluster.

Independent chip valuations illustrate both the opportunity and the sensitivity of the calculation. The Financial Times cited a forthcoming Barkr AI study valuing an eight-GPU H100 system from 2022 at roughly $320,000 today, about its initial price. Under a scenario in which compute supply catches up with demand, the study projects that system would retain about two-thirds of its value after another year and decline to around $30,000 after six years. The figures are a modeled path, not a guaranteed market price, but they give financiers a way to frame a loan’s duration, collateral coverage and premium. Barkr founder Thomas Galbraith said lenders need to understand both the income a chip can produce and its resale value. That data becomes more consequential as clusters are financed for longer periods.

Nvidia announced partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR in August to mobilize more than $500 billion of third-party capital for AI infrastructure over time. Insurance would add a distinct layer to that effort: investors can finance the project, while a risk-transfer structure can protect a lender against a sharp fall in collateral value. That may be especially useful to a neocloud trying to turn contracted demand into hardware capacity without paying for every system from equity. It also gives Nvidia a way to support purchases beyond the biggest technology companies without funding each buyer directly. The negotiations reveal how chip resale values, insurance pricing and cloud credit quality are becoming part of Nvidia’s commercial ecosystem alongside processors and software.