$NVDA Partners with Major Asset Managers on $500B AI Infrastructure Financing Framework

$NVDA has partnered with six of the world's largest asset managers—Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR—to mobilize more than $500 billion in third-party private credit and infrastructure financing to address a looming AI capital shortfall.
Bank of America estimates total artificial intelligence capital expenditures could surpass $5 trillion between 2026 and 2030, with roughly $1.2 trillion requiring external funding across capital-intensive operators such as $ORCL, SpaceX, CoreWeave, and Nebius. Under the proposed financing structure, private credit platforms purchase GPUs and lease them to end-customers, with $NVDA exploring residual value backstops of up to 25% to help lower borrowing costs.
The initiative has sparked debate across credit markets over chip collateral longevity and depreciation schedules:
• Useful-Life Divergence: CEO Jensen Huang argues enterprise GPUs can function as productive assets for up to a decade, whereas hyperscalers like Alphabet ($GOOGL) and $MSFT model four to six years, and investor Michael Burry projects an economic lifespan closer to two to three years.
• Secondary Value: Older models continue to demonstrate residual utility; six-year-old Nvidia A100 processors still trade near $5,000 each, shifting from frontier model training to secondary inference workloads.
• Guarantee Spreads: Market transactions reflect the premium placed on corporate backstops. A comparable $35 billion chip-backed facility for Broadcom priced at 5.75% with a corporate guarantee, compared to an estimated 8.5% unbacked rate—a 275-basis-point spread.
While the lending framework unlocks liquidity to sustain aggressive hardware orders, credit analysts caution that vendor guarantees effectively shift long-term depreciation, obsolescence, and technological execution risks back onto equipment manufacturers.