Nvidia mitigates major risk with hyperscaler commitments and $279 billion supply chain lockdown

For years, the knock on Nvidia was simple: too many eggs in too few baskets. When a handful of hyperscalers account for the majority of your revenue, one budget cut at Microsoft or a pivot at Google could send shockwaves through your entire business. Nvidia appears to have heard the criticism, and it’s responding with roughly $279 billion in multi-year supply commitments and a deliberate push to broaden who’s buying its chips.

Nvidia’s most recent quarterly earnings showed revenue of $96.2 billion, a 106% increase year-over-year. Data center revenue alone hit $89 billion.

The supply chain bet

Nvidia’s $279 billion in purchase agreements cover high-bandwidth memory, manufacturing capacity, and other critical components. The spending is spread across three fiscal years: $92 billion in the remainder of the current fiscal year, $87 billion in fiscal 2028, and $88 billion in fiscal 2029.

Back in 2025, the company disclosed a $6.3 billion take-or-pay commitment to CoreWeave for GPU capacity running through 2032. The current commitments dwarf that figure by orders of magnitude.

Diversifying beyond the big clouds

Nvidia’s quarterly results reveal a telling split in its data center business. Hyperscalers generated $48.7 billion in revenue. But Nvidia’s AI cloud and enterprise segment, which it calls ACIE, brought in $40.3 billion.

ACIE posted 138% year-over-year growth compared to 102% for hyperscalers. Hyperscalers still represent about 55% of data center revenue. ACIE encompasses AI labs, government entities, sovereign wealth-backed projects, and enterprises building their own AI infrastructure.

Following the money into AI financing

Nvidia has partnered with six major financial institutions to channel over $500 billion in third-party capital toward AI data center projects. Nvidia’s exposure in these arrangements is capped, with optional project-specific residual-value support limited to 25%.

Separately, Nvidia has committed up to $105 billion in guarantees supporting the development of an OpenAI-related campus in Ohio. That guarantee comes with provisions for reimbursement by OpenAI.

The risk hasn’t vanished entirely. A $279 billion commitment cuts both ways. If AI spending decelerates sharply, Nvidia could find itself locked into purchasing components it doesn’t need. The OpenAI guarantee, even with reimbursement provisions, represents meaningful contingent liability.