Nvidia taps Wall Street for a half-trillion dollars to fuel global AI infrastructure buildout

Some of Wall Street’s biggest financial firms are partnering with Nvidia Corp. to pour a half-trillion dollars of funding into the artificial intelligence industry’s massive infrastructure buildout.

Nvidia said today it has struck deals with Apollo Global Management Inc., BlackRock Inc., Blackstone Inc., Brookfield Corp., Goldman Sachs Group and KKR & Co. Inc. For the first time, those investors are treating AI hardware and infrastructure as an asset class like stocks, bonds and commodities, the chipmaker added.

“In AI, compute is revenue,” said Nvidia Chief Executive Jensen Huang. “We are bringing the world’s leading long-term capital providers together to independently underwrite AI infrastructure.”

Nvidia said the funds from today’s deals will go toward both its own projects and those of its partners. Some of the projects backed by the fund will include the construction of new data centers needed to house, operate and maintain servers filled with hundreds of thousands of Nvidia’s graphics processing units, which are widely used to process AI workloads. The money will also be used to back new manufacturing facilities to produce those chips in order to meet growing customer demand.

Nvidia has become the single largest beneficiary of the AI boom. These days, basically every major AI firm and technology company uses its chips to power AI features, services and chatbots. It could even be argued that basically every large organization in the world has indirectly become a customer of Nvidia’s, for few companies these days don’t use some form of AI tools in their day-to-day business operations.

Some of Nvidia’s biggest direct customers include Google LLC, Microsoft Corp., Meta Platforms Inc., Amazon.com Inc., SpaceX Corp., OpenAI Group PBC and Anthropic PBC. Collectively, these companies have spent more than a trillion dollars on AI projects and infrastructure in the last three years, and they’re expected to invest even more in future. A huge chunk of that money has, and will continue to find its way into Nvidia’s bank accounts, which is why the chipmaker’s stock has increased fivefold over that three-year period.

By using institutional credit, insurance funds and private capital to underwrite new AI infrastructure projects, Nvidia is helping its customers secure the financing they need without drawing on their own balance sheets. “This is really the first time that technology chips have become an investable asset class,” Huang told CNBC in an interview. “These are revenue-generating assets now. They’re productive, long-lived, fungible and flexible.”

Traditionally, GPUs have always been seen as depreciating investments that quickly lose value upon delivery to customers. But Nvidia is challenging that assumption, arguing that compute capacity is a longer-term and bankable asset class that’s widely adopted and transferable across customers. It means lenders can reliably underwrite compute as a revenue-generating asset, though skeptics may question if GPUs really retain their value when newer generations of the chips emerge.

“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 argued.

KKR co-CEOs Joe Bae and Scott Nuttall have certainly bought into Huang’s argument. They said in a joint statement that AI is already so pervasive and important that compute has become a critical asset. “As we’ve scaled our approach to digital infrastructure, we’ve learned that delivery, not ambition, is the hard part,” they added.

However, not everyone agrees with Nvidia. Some investors have become wary of the chipmaker’s alleged “circular dealmaking” involving eyewatering amounts of money, and today’s announcement will likely stoke those fears.

Last month, for instance, Nvidia announced a $500 billion deal with the South Korean semiconductor giant SK hynix Inc. to secure a supply of memory chips. Shortly after announcing that deal, reports emerged claiming that the company was discussing a $250 billion deal with OpenAI to help finance the AI giant’s massive 10-gigawatt data center project in Ohio, which is expected to be one of the world’s largest “AI factories” once it’s completed in 2028. No agreement has been confirmed so far, but if it is, it would represent one of the company’s biggest deals with a customer, Bloomberg reported. The $250 billion would only cover the data center lease and debt, and talks are ongoing regarding a separate, $350 billion deal to finance chip purchases, the report added.

Critics say Nvidia is weaving a dangerous, tangled web of deals with a handful of companies that have overlapping interests, many of which have struck multibillion-dollar deals with each other. The concern is that these financial dependencies mean that if one deal falls apart, it could lead to a domino effect that engulfs not only the AI industry, but the entire global economy, given how much money is at stake.

Photo: Nvidia