Nvidia has committed to cover up to 25% of losses on AI hardware that loses value, a guarantee that could reach $125 billion across the deals it is courting — and much of that risk is set to land on insurers' balance sheets.

On August 10, Nvidia announced memorandums of understanding with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to build financing platforms mobilizing more than $500 billion in third-party capital for AI infrastructure, Axios reported. The model is simple: raise money from pension funds, sovereign wealth funds and insurers; buy Nvidia chips and build data centers; lease the capacity to AI companies for years of steady payments.

To attract investors who do not normally underwrite chip depreciation, Nvidia added a sweetener. It said it "may provide a residual-value support mechanism for up to 25% of an opportunity, assessed carefully on a project-by-project basis." That covers part of the gap if hardware proves worth less than expected when a lease ends.

Where does the risk actually sit?

None of the $500 billion sits on Nvidia's own balance sheet, and neither does most of the credit risk behind it. Special purpose vehicles buy the GPUs and issue debt. Increasingly, CNBC reported in April 2026, that debt is placed with insurance and retirement capital managed by firms like Apollo and KKR.

The match is structural, not accidental. Data centers are decades-spanning projects, and they align with the long-dated liabilities insurers already carry — annuities and pension obligations. The risk does not disappear. It changes owners and arrives on an insurer's books relabeled as investment-grade fixed income.

CNBC's April reporting quotes the phrase "GPU debt treadmill," coined by AI commentator Dave Friedman, to describe the core mismatch. GPUs last roughly seven years. The data centers housing them are built for 20 to 30. Lenders underwriting a 20-year facility are betting today's chips, or their replacements, hold value in year fifteen — after the hardware has been swapped out two or three times.

How are insurers responding?

Capacity is a live problem. One insurance executive told CNBC that insuring a single $20 billion data center campus was nearly impossible to price in 2023. By 2026, in that executive's words, it has become a weekly conversation. Insurers have responded with new products:

• Cover for credit losses
• Cover for declines in chip resale value
• Cover for contract breaches tied to power outages or cooling failures

The debt underneath has grown fast. Incremental borrowing funded about 9% of hyperscaler capital spending in fiscal 2024. By mid-2026, that figure had climbed to roughly 32% on a trailing basis. U.S. data-center debt issuance roughly doubled to about $182 billion in 2025, per reporting Axios cited.

Why the timing matters?

Nvidia is not acting in a vacuum. Bain & Company said this week that the AI industry needs to generate $6 trillion a year in new revenue by 2031 to justify data center spending already committed. Existing AI services might cover $1.8 trillion, leaving a $4.2 trillion gap to fill from businesses that barely exist yet — autonomous robotics and AI-driven drug discovery among them. Bain separately projects $5 trillion to $6.5 trillion in data center spending through 2030.

Those numbers explain why insurers are being asked to carry so much of the weight. Traditional lenders and Nvidia's own balance sheet cannot absorb spending at that scale alone.

Nvidia's arrangement does not eliminate risk. It reroutes it — from Nvidia's income statement and from the banks that might otherwise hold this debt — into the reserves of companies whose business model depends on pricing tail risk decades in advance. If AI demand compounds as Nvidia's revenue guidance assumes, the residual-value guarantee mostly sits unused. But if a major AI lab pulls back capacity, or a cheaper chip generation strands the current fleet, the 25% backstop gets tested for real.

What is actually confirmed?

None of the six firms named in the MOUs have disclosed how much capital they have committed, or whether the residual-value terms are finalized. Nvidia's own release says the partnerships remain subject to final agreements.

The market has already reacted to $500 billion in financing infrastructure and a $125 billion guarantee ceiling. So far, not a single dollar of it is confirmed as spent.