The Big Short Investor Michael Burry Just Found a $537 Billion Problem With AI

Artificial intelligence has become Wall Street’s biggest growth story, with technology companies pouring hundreds of billions of dollars into chips, data centers, and computing power. Investors have largely accepted that spending as the price of admission to the next technological revolution. 

Yet the industry is increasingly building tomorrow’s infrastructure with money it expects to earn years from now. That leaves an uncomfortable possibility: AI could transform the economy without generating enough profits to justify everything being spent on it. And one investor famous for spotting financial disasters before they happen believes he’s found evidence of that problem hiding in plain sight.

Michael Burry Finds a $537 Billion Warning

Michael Burry, whose bet against the housing market became famous in The Big Short, has identified what he considers a dangerous financial imbalance in AI developer Anthropic‘s newly released initial public offering prospectus.

According to Burry’s Oct. 27 commentary, Anthropic has accumulated approximately $537 billion in computing commitments across multiple suppliers, including:

^{Source: Michael Burry’s analysis of Anthropic’s IPO disclosures. Figures include estimates and conditional arrangements, not solely fixed, immediately payable liabilities.}

Anthropic’s Oct. 26 SEC filing separately identifies approximately $417 billion in future minimum payment obligations under additional compute and infrastructure agreements.

The difference between them is important. Burry’s $537 billion isn’t a conventional debt balance, nor is every dollar guaranteed to be spent. But both figures reveal how much future business Anthropic must generate to support its expansion. And the implications extend well beyond one AI developer.

The Revenue Problem Hiding Behind AI’s Growth

Anthropic reported $16.3 billion in revenue during the first half of 2026, up approximately 1,300% from the previous year. Its operating loss narrowed to $693 million, compared with $2.1 billion a year earlier. Those figures hardly describe a failing business.

The trouble is that approximately 80% of Anthropic’s revenue comes from consumption-based services. Customers pay for the computing they use, often without making comparable long-term commitments. Anthropic, meanwhile, must pay for infrastructure whether customers use it or not.

Its business might be booming today, but tomorrow’s rent doesn’t disappear if sales suddenly slow. Burry worries that the massive AI infrastructure spending cycle has created exactly this mismatch.

Compounding the problem, increasingly efficient AI models could reduce the computing power needed for each task. Lower costs might encourage greater usage, but they could also weaken demand for expensive infrastructure.

Why Nvidia and Other AI Stocks Are Exposed

Anthropic isn’t shouldering these risks alone. Amazon, Microsoft, and Alphabet provide computing capacity. Broadcom supplies specialized chips and helps finance infrastructure. Nvidia (NASDAQ:NVDA) benefits from the broader demand for AI processors.

These companies have different levels of exposure, but they share one vulnerability: their investment plans increasingly depend on AI demand continuing to grow. As Broadcom’s expanding financing arrangements demonstrate, AI infrastructure has become intertwined with loans, leases, and long-term purchasing contracts.

If Anthropic’s growth slows, suppliers might face weaker orders, lower equipment utilization, or more difficult financing conditions. Admittedly, Anthropic’s revenue growth is real, and its operating losses have narrowed. Its suppliers also serve customers beyond Anthropic, providing some protection.

Still, Google’s growing off-balance-sheet commitments illustrate how financial exposure can extend beyond traditional borrowing. The bigger danger isn’t that AI disappears. It’s that investors discover the infrastructure supporting it was priced for more demand than ultimately materializes.

Key Takeaway

Michael Burry hasn’t proved the AI bubble is about to burst. His $537 billion calculation does, however, expose a vulnerability that investors shouldn’t dismiss. AI developers are making multiyear financial promises against revenue that customers can reduce almost overnight.

For investors, the answer isn’t necessarily to abandon Nvidia, Microsoft, or other AI leaders. It’s to prioritize companies generating free cash flow, monitor how quickly infrastructure spending translates into profits, and avoid paying valuations that assume uninterrupted growth.

Ultimately, AI can change the world while still producing disappointing investment returns. Burry’s warning is a reminder that the two aren’t mutually exclusive.

Contact [email protected] for any questions or corrections.