NVIDIA Limits Exposure to $125 Billion in New $500 Billion Compute Fund
NVIDIA has structured a $500 billion compute fund with a 25 percent backstop to finance second-tier AI operators, a move designed to diversify revenue away from hyperscalers but raising structured credit concerns.
NVIDIA has unveiled a $500 billion investment partnership aimed at financing artificial intelligence infrastructure. The chipmaker will cap its direct exposure at $125 billion through a 25 percent backstop. The remaining capital will be housed within special purpose vehicles.
These special purpose vehicles are legal entities designed to hold specific assets and debt separate from NVIDIA’s balance sheet. Lenders will finance the entities, which in turn will own the graphics processing units and data center capacity. NeoCloud operators will then lease this compute power.
The 25 percent backstop functions as a first-loss cushion. This structure is intended to provide senior lenders with sufficient comfort to fund the remainder of the vehicle.
The primary strategic objective behind this financial engineering is diversification. Hyperscale clients currently account for approximately half of NVIDIA’s total business, a heavy concentration detailed in the company’s own filings.
In the first quarter of fiscal year 2027, NVIDIA reported Data Center revenue of $75.25 billion. This figure represents a 92 percent year-over-year increase, with hyperscale customers contributing roughly 50 percent of that total.
By financing second-tier buyers, NVIDIA aims to extend the demand curve. This ensures continued growth even if capital expenditure from major hyperscalers begins to plateau. The fund is already targeting NeoCloud operators such as Nebius Group, which recorded 454 percent revenue growth in the last quarter.
However, the structure has drawn scrutiny from market observers. Austin Smith recently highlighted a comparison made by the Goldman Sachs chief executive, who likened the fund’s architecture to mortgage-backed securities.
This comparison serves as an inadvertent warning regarding concentration risk within special purpose vehicles. Investors will now watch closely to see if this model creates a durable financing flywheel or introduces systemic vulnerabilities into the AI buildout.
