Nvidia, Wall Street Firms Seek $500 Billion to Turn GPU Chips Into Financeable Assets

Letting GPU-hungry companies borrow against computing capacity

Third-party financiers join after criticism of 'AI circular finance'

Nvidia, Wall Street Firms Seek $500 Billion to Turn GPU Chips Into Financeable Assets

Nvidia is teaming up with six major Wall Street asset managers on a $500 billion fundraising effort aimed at turning artificial intelligence chips into a new asset class. The plan would treat graphics processing units, or GPUs, as collateral with financeable value, similar to commercial real estate, toll roads and other infrastructure assets, allowing companies to borrow against them.

The initiative effectively puts Wall Street firms in charge of supplying capital to Nvidia customers using GPU chips as collateral, after criticism that Nvidia was engaged in "AI circular finance" by investing in companies that buy its chips.

CNBC reported on August 10 that Nvidia plans to build a financing platform for its customers with BlackRock, Apollo Global Management, Blackstone, Field Asset Management, Goldman Sachs and KKR.

Executives from the seven companies also appeared together on CNBC that day to explain the effort.

The goal is to bring in outside capital so Nvidia customers, including hyperscalers, advanced AI labs and corporations, can build data centers and purchase Nvidia hardware.

The structure could reshape how AI infrastructure is funded. Nvidia would tap institutional credit, insurance funds and private capital to support investment in GPUs and data centers, helping end users obtain financing without committing their own cash.

Chief Executive Officer Jensen Huang told CNBC it was the first time technology chips had become investable assets. Those chips now generate income, he said, describing them as productive, long-lived, transferable and flexible.

Huang argued that Nvidia's AI chips are widely used and can be transferred between customers, giving financial firms grounds to underwrite them as long-duration, income-producing assets.

Historically, GPUs have been viewed as hardware that loses value quickly. Nvidia's push amounts to an attempt to turn AI computing capacity into long-term, profitable infrastructure. Skeptics, however, may question whether those assets can retain their value as next-generation AI chips arrive.

Huang told CNBC the fundamental shift in the industry is that computers have now become part of infrastructure, much like electricity or the internet. Computing should therefore be treated as infrastructure, he said.

Alternative asset managers have already been active in channeling institutional and insurance capital into digital infrastructure projects. Apollo, Blackstone and others have provided debt and equity financing to a range of companies, including Anthropic.

Since the AI-led rally cooled in July, investors have questioned and criticized AI circular finance. As hyperscalers spend hundreds of billions of dollars on data centers and hardware, ratings firms such as Moody's have warned that the unprecedented capital expenditures are squeezing cash flow and adding to debt burdens at big technology companies.

BlackRock, Goldman Sachs, Blackstone and other major Wall Street firms said in a release that computing technology has rapidly established itself as a core asset that will drive the next wave of global economic growth.

Goldman Sachs Chief Executive Officer David Solomon said in the release that the firm's participation reflects its confidence in Nvidia's leadership. Goldman also sees strong potential in the new opportunity to create a credit market built on Nvidia-powered computing, he said.

Huang also told CNBC that he personally pitched the idea of computing-based finance to Wall Street heavyweights.

Blackstone President Jon Gray told CNBC that AI computing would come to be viewed as a financeable asset class, much as mortgage lenders view housing. Demand for AI is outstripping supply, he said, adding that AI usage at Blackstone portfolio companies has surged sevenfold this year.

BlackRock Chief Executive Officer Larry Fink said he believes the project marks the start of "the next future of financial engineering," comparing it to the birth of mortgage-backed securities in the 1970s. Some money has already been raised, and BlackRock expects to raise substantially more, he said. Fink also stressed that it is critical for the US to become the global leader in AI.

Kim Jung-a, guest reporter, Hankyung.com, kja@hankyung.com