Nvidia and six Wall Street giants are turning GPUs into a $500 billion asset class

Nvidia signed memorandums of understanding with six of the largest financial institutions in the world on Monday, creating dedicated financing platforms to raise more than $500 billion for AI infrastructure.

Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR are the partners.

The platforms are designed to let Nvidia's customers, including hyperscalers, frontier AI labs, and enterprises, borrow against compute purchases the way a company borrows against a building. Nvidia's GPUs and full-stack AI infrastructure become the collateral.

I think the word "memorandum" is doing a lot of work here. These are agreements to agree. The final terms, individual commitments, and a timeline for deploying the money were not disclosed.

But the framing was deliberate. Every executive quoted in the announcement used the same language: compute is infrastructure now.

"This is really the first time that technology chips have become an investable asset class," Jensen Huang, Nvidia's founder and CEO, told CNBC. "These are revenue-generating assets now. They're productive, they're long-lived, they're fungible, they're flexible."

"Fundamentally, what's different about this industry and this way of doing computing is that the computer is now part of the infrastructure, like electricity, like the internet, and so you have to think about it like it's infrastructure," Huang said.

Nvidia said it has the option to backstop up to $125 billion of the potential deals, or 25%, a number Huang disclosed on X.

David Solomon, Goldman Sachs' CEO, said in the announcement that compute will become a "financeable asset class" in the same way mortgage lenders look at homes. Jon Gray, Blackstone's president and COO, made the same comparison on CNBC. Larry Fink, BlackRock's CEO, went further, calling the project "the next future for financial engineering" and comparing it explicitly to the creation of mortgage-backed securities in the 1970s.

"We need to raise this money as fast as possible and put this to work, because I think it's really imperative that the United States is the leader in AI in the world," Fink said on CNBC.

Huang came up with the idea and pitched the Wall Street firms himself, Solomon said. Jensen calling on Wall Street to create a credit market for his chips is not subtle. Nvidia's customers have been spending unprecedented amounts on AI infrastructure, and those bills are starting to strain balance sheets.

Big Tech companies are on track to spend more than $730 billion on AI this year combined. Moody's warned in July that the capital expenditure levels are starting to squeeze free cash flow and push tech giants into heavier debt loads. The financing platforms solve a simple problem: Nvidia's customers need the chips, and they need someone else to front the money.

The circular nature of AI financing has drawn scrutiny. AI companies invest in other AI companies on the condition that those companies buy their products. Nvidia financing its own customers closes the loop tighter.

The compute lives in a data center, generates revenue, services the debt. The lender gets usage-linked returns on a long-duration asset.

"Skeptics may question whether AI chips can retain their value as newer generations emerge," CNBC reported in its coverage. That is the open question, and it is not a small one.

Chips depreciate. The RTX 4090 I bought two years ago is worth maybe half what I paid. Nvidia is arguing that data center GPUs are different because demand is so far ahead of supply that even last-generation chips stay utilized.

I get the sense this works until it does not. The mortgage-backed securities market also seemed airtight until the underlying assumptions about housing prices broke.

Compute-backed securities rest on the assumption that AI demand keeps outpacing chip supply, that Nvidia keeps shipping new architectures that are backward-compatible, and that the offtakers keep offtaking. Any one of those assumptions could shift.

Huang addressed this head-on in the CNBC interview, arguing that Nvidia's software ecosystem, specifically CUDA, extends the useful life of its hardware and improves its economics over time.

Blackstone's Gray said demand for AI is outstripping supply, with AI use at Blackstone portfolio companies surging sevenfold this year. That is the kind of number that makes you sign an MOU.

Nvidia's stock closed down 2.4% on the day, which surprised me. A $500 billion financing backstop for your core product is the kind of announcement that usually moves a stock up. Investors seemed to be pricing in the execution risk, or maybe the dilution from the backstop commitment.

The comparison to the mortgage market matters. When Fink says this is MBS 2.0, he is describing the mechanics, but people heard the risks.

Securitization created massive liquidity and massive systemic fragility. Compute-backed securities are starting from a different place: the underlying asset actually produces revenue, and demand is genuine. But the financial engineering has the same shape.

I am not sure whether to call this a financing arrangement or a category creation event. The signing of MOUs with the six largest names in alternative asset management tells me the category already exists and the infrastructure is just catching up.

When BlackRock, Blackstone, Apollo, KKR, Goldman, and Brookfield all agree something is an asset class, it probably is one. The question is whether the asset holds its value across chip generations, and nobody has an answer to that yet.

For any company spending on AI infrastructure, the financing window just opened. Nvidia's competitors will not sit still. If compute-backed lending becomes the standard, the finance layer becomes as important as the silicon layer, and Nvidia just locked in the six largest lenders in the world before AMD or anyone else could pick up the phone.

Source