Chip giant and six finance titans target over $500 billion to fund “AI factories.”
Nvidia just moved beyond selling the picks and shovels of the AI boom. On Monday it announced partnerships with six of Wall Street’s heaviest hitters to mobilize more than $500 billion in third-party capital. This money will go to the data centers and computing power that train and run AI models.
The chipmaker signed memorandums of understanding with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR. The goal is to create independent “compute financing platforms” that treat Nvidia hardware and full-stack AI infrastructure as an investable asset class. As a result, lenders can underwrite these like real estate or energy projects.
Why compute became collateral
Demand for AI capacity keeps outstripping supply. Hyperscalers, frontier labs, enterprises and governments need massive clusters of GPUs, power and cooling. Paying for them upfront strains balance sheets. The new platforms aim to supply capital at scale and attractive rates so customers can finance the buildout instead.
Jensen Huang framed the shift clearly: “We began by building chips; today, we are helping create a new class of productive, investable infrastructure: AI factories. In AI, compute is revenue.” He argued Nvidia’s systems are broadly adopted, flexible across workloads, transferable and improved over time by CUDA software. Because of this, they have a longer economic life than typical tech hardware.
Nvidia said it may backstop residual value on up to 25% of opportunities—roughly $125 billion if the full target is reached. However, it stressed that the six firms will independently underwrite each deal based on customer demand, utilization and cash flow.
Market reaction and bigger picture
Nvidia shares fell on the news, dropping as much as 3% and ending the day lower. Some investors worry about circular financing loops in the AI buildout. Yet Huang pushed back, saying the structure brings independent long-term capital into the market. He also said demand is real, coming from labs, startups, enterprises, cloud providers and countries.
Executives from the partners sounded bullish. Goldman Sachs CEO David Solomon called it a pivotal moment in a historic AI investment cycle. He said the firm is excited to help create a market for credit backed by Nvidia compute. BlackRock’s Larry Fink highlighted the need for both capital and skilled workers to turn investment into productive infrastructure.
The agreements remain subject to final terms. No specific projects or timelines were named. The $500 billion figure is an aggregate target for capital raised over time, not Nvidia revenue or a single fund.
What it means for the industry
If the platforms work, smaller AI players and enterprises could gain easier access to scarce compute without exhausting their own cash. For Nvidia it reinforces the company’s central role: not just selling chips, but helping the entire ecosystem finance the factories that use them. For Wall Street it opens a new asset class tied to one of the fastest-growing technologies on the planet.
Whether the hardware’s useful life and residual value hold up under heavy, continuous AI workloads will determine if this financing model scales or hits limits. For now, Nvidia has lined up some of the deepest pockets in finance to test the idea.