Nvidia has signed MoUs for $500 billion of private credit and private equity from six major Wall Street firms: Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR.

BlackRock CEO Larry Fink compared this initiative to the birth of mortgage-backed securities (MBS) in the 1970s-80s, but this time with computer chips and data centers holding the potential for long-term revenue-generating infrastructure. While millions of Americans remember MBS as a financial crisis that hurt millions of families, Fink’s lens as one of the MBS pioneers differs, as MBS was a financial engineering mechanism that unlocked massive pools of global capital so that homeownership has become more accessible to tens-of-millions of Americans (today, with stricter regulation, the U.S. MBS market spans well over $11 trillion).

In the AI era, the banks look at a heavy-duty industrial asset class as an opportunity to widen the pool of capital, with infrastructure investors, private-credit providers, banks, and asset managers bringing a scale of funding not possible through individual tech firms. In their eyes, larger-scale financing will feed the entire ecosystem around AI hardware, such as  electricity grids, gas-fired and renewable generation, transmission capacity, and land for data centers. 

On the other hand, critics, short sellers, and economic watchdogs are raising alarms about the systemic vulnerability of providing debt to unprofitable startups and tech companies that will purchase more chips from Nvidia. Like the homeowners in the MBS crisis, some critics say ordinary citizens – whose retirement savings are managed by the institutional investors funding this deal – could ultimately be harmed, as could institutional asset managers, private-credit investors, tech clients/startups, and Nvidia itself.

With Nvidia providing a financial backstop for up to 25% ($125 billion) of the project costs, critics say the company will be deeply exposed if end-users default. Tech analyst Ed Zitron called the structure “an insane thing to do on so many levels.”

For now, if it goes through, the deal will most likely act as a massive accelerator for AI, but, with very high-stakes risks. There will be winners and losers.