On August 10, NVIDIA announced it had signed memorandums of understanding with six major financial institutions—Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR—to establish independent financing platforms. The initiative aims to mobilize over $500 billion in third-party capital over time for AI infrastructure development. NVIDIA CEO Jensen Huang called this a milestone, marking the company's evolution "from building chips to helping create a new type of productive, investable infrastructure: AI factories".
Model Innovation: Turning Compute into "Digital Real Estate"
The financing structure's core is to position AI compute infrastructure as a collateralizable asset, akin to commercial real estate or toll roads. The six institutions will create large-scale capital pools to provide financing at competitive rates to NVIDIA's ecosystem clients, including AI labs and cloud providers. Clients can use these funds to procure hardware and build AI factories without depleting their own balance sheets, while capital providers gain long-term returns tied to actual compute usage.
Huang defined NVIDIA compute as a new "investment-grade asset," citing its widespread adoption, flexibility across models and workloads, transferability between customers, and continuous value improvement through CUDA software. He noted the A100, launched in 2020, remains in active commercial use, with its economic lifespan extending toward a decade.
Market Reaction and Concerns: Huang Defends the Model
Despite the high-profile partnership, market response was cautious. NVIDIA's stock fell approximately 2.9% on the announcement, erasing over $70 billion in market value. Its five-year credit default swaps (CDS) posted their largest one-day increase in nearly two weeks.
The primary concern is "circular financing"—NVIDIA financing clients who use the funds to buy NVIDIA hardware, creating a self-reinforcing loop. "Big Short" investor Michael Burry publicly warned this could push circular spending to unprecedented scales.
Huang responded across official statements, media interviews, and social media. He emphasized that real demand comes from AI labs, startups, enterprises, and cloud providers, and that capital providers will independently underwrite each project. He clarified the $500 billion figure represents total third-party capital the platforms are designed to mobilize, "not NVIDIA revenue, nor a single fund or commitment to one customer".
Industry Context: The Funding Gap and Capital Influx
The financing platform arrives amid a massive funding gap for global AI infrastructure. Morgan Stanley projects hyperscalers will spend approximately $3.5 trillion on capital expenditures between 2026 and 2028. Apollo president Jim Zelter estimated total AI infrastructure investment could exceed $8 trillion.
For NVIDIA, this unlocks external leverage beyond its own balance sheet; for clients, it alleviates the upfront cost burden of acquiring large-scale GPU clusters. Goldman Sachs CEO David Solomon revealed the idea originated with Jensen Huang. BlackRock CEO Larry Fink likened it to the dawn of mortgage-backed securities, calling it the "beginning of a new chapter in financial engineering".
NVIDIA's move essentially repackages AI compute—traditionally seen as rapidly depreciating hardware—into a new asset class with stable cash flow, leveraging Wall Street's long-term capital. While this addresses real global compute shortages and secures future demand, the key variable remains whether genuine AI revenue can support the massive financial leverage being deployed.