Nvidia taps six Wall Street giants in bid to unlock $500 billion for AI infrastructure

Nvidia has a problem most companies would love to have. Demand for its AI chips keeps climbing, but plenty of would-be buyers can’t comfortably pay for them.

Its answer is a new financing machine. On August 10, 2026, Nvidia announced memorandums of understanding with six of the biggest names in finance: Apollo Global Management, BlackRock, Blackstone, Brookfield Asset Management, Goldman Sachs, and KKR.

The goal is to build independent “compute financing platforms” that aim to raise more than $500 billion in third-party capital for AI infrastructure. That money would fund GPU purchases and data center construction.

How the financing model works

The initiative frames AI hardware as a revenue-generating asset class, comparable to aircraft leasing or utilities. Rather than writing a giant check upfront, customers could tap outside capital to get access to Nvidia hardware. The target audience includes AI labs, cloud providers, and enterprises squeezed by capital constraints in a high-interest-rate environment.

Nvidia itself isn’t putting cash in or taking on new debt. The company may provide residual-value guarantees of up to 25% on certain projects. Across the full program, that could translate to a maximum of approximately $125 billion in support.

Wall Street’s skepticism

As of October 1, 2026, Wall Street investors had voiced concerns about whether chips can serve as reliable long-term collateral. The worry centered on guarantees that were seen as insufficient. AI hardware evolves quickly, and today’s flagship GPU can look dated sooner than a lender holding a multi-year loan would like.

The expectation is that forthcoming agreements will include stronger protections and revenue backing. If that happens, the overall effort is viewed as credit-positive for Nvidia.

The CoreWeave blueprint

Nvidia isn’t inventing GPU-backed lending from scratch. CoreWeave launched an $8.5 billion investment-grade facility backed by GPUs. That deal showed lenders were willing to underwrite large projects secured by compute hardware.

Nvidia’s plan scales the concept dramatically. The $500 billion target is roughly 59 times the size of the CoreWeave facility, and it involves six institutional heavyweights instead of a single borrower.

CEO Jensen Huang has framed the partnerships as doing two jobs at once. In his telling, they keep hardware demand growing while also giving institutional investors, such as pension funds and sovereign-wealth funds, a way into AI infrastructure with reduced risk exposure.

What this means for Nvidia and AI buyers

Nvidia’s potential exposure of approximately $125 billion is contingent, not cash out the door, but it’s still a meaningful commitment tied to how well GPUs hold their value.

Key signals to track include whether the MOUs convert into binding agreements and how much capital actually gets raised against the $500 billion target. The terms of those deals, especially the revenue backing and collateral protections that skeptics flagged, will show whether GPUs can truly be underwritten like aircraft or utilities.