Nvidia teams with major financiers to package GPU compute as a new asset class
Nvidia announced a partnership with several large investment firms to create financing structures that treat its GPU compute as an investable asset.
Nvidia disclosed that a group of leading investors - Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR - will collaborate on a financing program valued at roughly $500 billion to classify GPU compute as an asset class. Jensen Huang told one outlet that the chips, together with Nvidia’s CUDA software, generate ongoing revenue and can be used as collateral for loans. The initiative is currently limited to memorandums of understanding, with no final contracts signed yet.
Huang cited the continued commercial use of older A100 GPUs as evidence that compute can retain value over many years. Analysts note that the success of the model depends on sustained demand for AI inference and the financial health of cloud providers that would borrow against the hardware.
Why it matters
It could reshape how AI hardware is financed, affecting tech companies and investors.
How this story developed
- Aug 10 Nvidia talks $500 billion AI infrastructure funding with major Wall Street firms
- Aug 11 Nvidia moved from talks to announcing memorandums of understanding with six firms.
- Aug 13 Nvidia said the financing will be structured as debt instruments that use compute capacity as collateral.
- Aug 14 Nvidia announced up to 25 percent residual‑value support on select GPU financing deals.
- Aug 17 Nvidia announced a financing program with six Wall Street firms that could mobilize more than $500 billion for AI data‑center projects.
In this story
Related stories
11 in this thread