Goldman Sachs in talks with investors on Nvidia’s $500 billion AI financing deal that could turn compute into Wall Street’s next asset class

Goldman Sachs is talking with banks, insurers, asset managers and private credit firms about backing Nvidia’s ambitious $500 billion AI financing initiative, according to Reuters, putting Wall Street deeper into a spending boom that could reshape how AI infrastructure gets funded.

The report comes just three days after TechStartups reported that Nvidia was turning to Wall Street for a $500 billion AI infrastructure financing push. The chip giant teamed up with Apollo Global Management, Blackstone, BlackRock, Brookfield Asset Management, Goldman Sachs and KKR to build financing platforms capable of directing more than $500 billion into AI infrastructure. Now, we’re getting a clearer picture of how that enormous financing machine could actually work.

The bigger story may be what Nvidia and its financial partners are trying to create: a market where AI compute can be financed, packaged and traded more like a traditional financial asset.

Nvidia unveiled the initiative on August 10 alongside six major financial institutions, with plans to mobilize more than $500 billion in third-party capital for AI infrastructure. Goldman has secured a central role as the sole lender, working alongside alternative asset managers including Blackstone and Apollo.

The scale is striking, yet it reflects an even larger capital problem confronting the AI industry. Goldman Sachs Research estimates that the four largest hyperscalers plan to spend more than $5 trillion on technology and data centers through 2030.

Corporate balance sheets alone may not be the most efficient way to finance that buildout. Wall Street sees an opening.

Goldman could provide junior capital and private credit financing through its asset management business. Its investment banking arm could place debt with private credit funds and eventually public debt investors, Reuters reported.

That would bring pools of institutional capital into an AI infrastructure market largely funded by technology companies, hyperscalers and specialized infrastructure investors.

From GPUs to a new financial asset

What makes Nvidia’s plan unusual is how it could distribute risk.

Previous AI infrastructure financings have leaned heavily on guarantees from technology vendors. Broadcom, for example, provided a residual-value guarantee connected to roughly $30 billion of senior debt supporting Anthropic’s AI chip financing.

Nvidia is taking a different approach.

CEO Jensen Huang said on X that Nvidia can backstop up to $125 billion, equal to 25% of potential deals. Much of the remaining risk would sit with the financing consortium rather than Nvidia itself.

“This appears to be a pivot away from vendor-financing,” Bank of America analyst Vivek Arya wrote in a note. “The burden sits with the consortium, not (Nvidia’s) balance sheet.”

The ambition goes further than funding another generation of data centers. The proposed structure could help establish an asset-backed market for AI compute, allowing debt linked to GPUs and computing infrastructure to trade more like conventional securities. A functioning secondary market could lower financing costs and open AI infrastructure to a far larger pool of institutional money.

That idea matters even more as Nvidia sits at the center of the AI economy. The company is now valued at roughly $5.2 trillion, making it the most valuable publicly traded U.S. company. Its chips have become core infrastructure for training and running many of the largest AI systems.

Goldman’s position in the deal stems from years of ties with Nvidia. The bank advised Nvidia on its $6.9 billion acquisition of Mellanox Technologies in 2019 and was among the lead underwriters on Nvidia’s $25 billion bond sale in June. Goldman CEO David Solomon has maintained a relationship with Huang, interviewing him at the bank’s technology conference less than two years ago.

“Jensen came, approached us with the idea, and we said we’d love to talk to you about it,” Solomon told CNBC this week.

If Nvidia and its partners succeed, the consequences could reach far beyond one $500 billion financing program.

For decades, Wall Street built enormous markets around assets that produced predictable economic value, from mortgages and aircraft to commercial property and infrastructure. AI compute could be entering that club.

GPUs started as hardware.

They became the scarce commodity behind the AI boom.

Now Wall Street is trying to figure out how to turn them into an asset class.