Nvidia has partnered with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to establish financing platforms targeting over $500 billion in third-party capital for AI compute infrastructure. The partnerships rely on a novel model: treating Nvidia hardware as a revenue-generating asset that lenders can underwrite directly.
The mechanism hinges on Nvidia CEO Jensen Huang's argument that the company's chips qualify as "investable assets" because they are broadly adopted, flexible and transferable between data centers. Under this framework, lenders evaluate compute capacity itself as collateral, not just the operator's creditworthiness. A borrower can finance Nvidia chips based on the expected cash flow from renting compute time to third parties, similar to how a bank might finance an industrial asset. The $500 billion figure represents a mobilization target across multiple signed memoranda of understanding rather than committed capital.
Nvidia has long dominated AI chip supply, capturing roughly 88% of the discrete GPU market for AI training as of mid-2024. Major cloud providers and independent data center operators have faced constraints in deploying new clusters fast enough to meet demand. This gap has created an opening for alternative capital structures.
Goldman Sachs, KKR, Blackstone and Apollo have each built infrastructure financing and lending practices over the past decade, often working on renewable energy and telecommunications assets. BlackRock manages over $10 trillion in assets globally and has moved into private credit. Brookfield operates one of the world's largest data center platforms and holds significant real estate and infrastructure portfolios. These firms are entering compute financing alongside their other asset classes.
The arrangement sidesteps traditional IT equipment leasing, where hardware rapidly depreciates. Instead, lenders are betting that Nvidia chips retain utility and revenue potential across multiple operators and use cases. This works only if Nvidia maintains its manufacturing lead and chips remain compatible across customer installations. A shift in Nvidia's roadmap or loss of market share would erode the collateral thesis underlying these financings.
The timing follows a wave of billion-dollar data center buildouts by Meta, OpenAI, Anthropic and others. These companies have struggled to secure enough chips to execute their plans, and third-party financing platforms could accelerate deployment by reducing the capital burden on individual operators. Goldman Sachs and KKR have both led recent rounds in AI infrastructure companies.
The platform structure also offers Nvidia an indirect route to expand addressable market without taking balance sheet risk. If third-party capital finances customer purchases, Nvidia's revenue can scale beyond what traditional enterprise credit alone would support. The company reported $60.9 billion in revenue for fiscal 2025, with data center accounting for 86% of sales.
Investor appetite for asset-backed AI compute structures will depend on demonstrable cash flows from operating data centers. The first wave of these financings will set terms and underwriting standards for the broader market. If default rates spike or compute capacity sits idle, institutional lenders may tighten terms or retreat. The number that decides whether this model scales is the actual revenue generated from financed clusters over the next 18 months.
