Tech Giants Digging Deeper Debt Hole To Reach AI Dreams

Nvidia and a group of major investment firms announced Monday they were launching a $500 billion financing initiative to fund new artificial intelligence infrastructure, which could revive jitters about a debt-induced bubble in the AI market spilling over into the broader economy.

The investment is the latest in a series of AI-induced debt financing rounds that have sparked fears about a potential AI bubble that could burst and spread to the rest of the economy. The funding group included BlackRock, Apollo Global Management, Blackstone, Brookfield Asset Management, Goldman Sachs and KKR, according to The Wall Street Journal.

Together they will issue loans and sell debt-backed investments to help tech companies build giant data centers and buy specialized computer chips. 

During a CNBC panel featuring other major CEOs involved in the deal, BlackRock CEO Larry Fink framed the effort as essential to keep the AI boom from running out of money, telling investors that raising cash quickly represents the future for financial engineering.

Nvidia had previously faced criticism for funding AI startups that used their loans to buy Nvidia chips—a practice critics described as “circular financing,” according to The Financial Times.

The $500 billion debt pipeline came as AI spending reached record levels. Tech companies and cloud providers were on track to spend $3.5 trillion on AI infrastructure through 2028, according to Morgan Stanley analysts. Recent deals highlight the eye-popping costs, such as AI developer Anthropic agreeing to lease $9 billion worth of computing power from Riot Platforms, according to CNBC.

To fund this massive push, tech companies and investment firms are turning to complex financial products similar to those that drove the 2007–2009 subprime housing crash as reported by The New York Times’ Dealbook newsletter in November 2025. Instead of home mortgages, these new packages bundle together high-risk tech loans, computer hardware leases and data center real estate, then sell them to institutional investors like pension funds.

Financial experts warn that bundling loans this way carries severe risks if the underlying computer chips quickly become outdated or if AI products fail to bring in enough cash. “When things go bad, they go really bad,” Dan McNamara, founder of Polpo Capital, told DealBook.

So far, everyday consumer spending on AI tools remains relatively small, generating about $12 billion annually, according to venture capital firm Menlo Ventures. Because of that gap, central banks—including the Bank of England—warn that relying heavily on debt to build AI projects creates broader risks for the global economy.

If AI fails to bring in enough profits to cover these loans, the financial fallout won’t just hit tech companies, it could spill over into the private credit markets and everyday pension funds that bought up the tech debt, market analysts at Bloomberg warned. This in turn could cause inflation for everyday Americans.

Supporters claim AI is essential for keeping the U.S. competitive globally. Industry leaders say building AI infrastructure rapidly is a national security priority to stay ahead of rivals like China, while promising future breakthroughs in medicine and overall job productivity.

But the rapid AI construction is hitting roadblocks across several states, with the rush to build data centers triggering pushback from local residents and city officials who fear the facilities will drain power grids and consume huge amounts of water.