Michael Burry, the investor who famously predicted the 2008 subprime mortgage crisis, has increased his bearish bet against NVIDIA (NVDA). In a disclosure on Substack, Burry revealed he has expanded his short position in the chipmaker as of August 11, and he openly criticized a $500 billion Wall Street financing deal for AI infrastructure announced this week, branding it a "Wall Street gimmick" with "shades of Enron-style operations."

Why the $500 Billion Deal Raises Red Flags

This week, Wall Street announced a plan to provide up to $500 billion in funding to NVIDIA's clients to expand AI infrastructure. Burry views this unprecedented move as a new form of "financial engineering." He warns that using structured credit to artificially prolong the bull market's momentum creates deep unease. Burry pointed out that the current scale of off-balance-sheet liabilities now dwarfs what was seen during the 2000 tech bubble and the 2008 financial crisis. He emphasized, "Structured credit is a natural part of the system, but artificially constructing unnatural credit in the late stages of a bull market to sustain momentum is what is truly concerning."

Circular Financing and 'Creating Its Own Demand'

A core risk Burry has long focused on is the circular financing model between NVIDIA and its customers: the company provides guarantees or financing to clients, who then use those funds to purchase NVIDIA chips. Meanwhile, NVIDIA also supports its own business by holding equity in clients and offering residual value guarantees. In a post on X, he noted that in this $500 billion deal, NVIDIA will hold approximately 25% of the equity and provide residual value guarantees on chip purchases, with the entire process filtered through private credit mechanisms from private equity. Burry stated bluntly, "I have a clear picture in my mind of what this will ultimately look like." In his view, this arrangement, while seemingly introducing external private credit, further complicates the industry's subsidy system. The AI sector is currently almost entirely debt-driven, and NVIDIA is using these methods to "create demand for its own products," thereby inflating reported revenue—while hiding massive future contingent liabilities.

Comparison to the Enron Scandal

Burry compared the current situation to the 2001 Enron bankruptcy. Enron, through complex accounting maneuvers, turned its commodities trading into an "investable asset class," eventually collapsing from a peak of $25 billion in daily trading volume. He urges investors to revisit Enron's history, arguing that "history is repeating itself." He suggests that not only is the AI bubble forming in a similar way, but the accounting methods used in these massive deals also make balance sheets appear more palatable. Burry noted, "There is no real mystery behind why the revenue numbers look so impressive—even as losses widen, free cash flow turns negative, and off-balance-sheet leverage explodes." Previously, an investment manager told Bloomberg that capital is increasingly being used to finance future AI customers and infrastructure deployment, a phenomenon that is itself unsettling. Burry also shared a chart showing that NVIDIA's approximately $5 trillion valuation is largely tied to these circular AI contracts.

Potential Impact on Markets and the Economy

Burry believes these operations are not just about NVIDIA alone, but concern the very foundation of the entire stock market and economy. The current AI boom is heavily debt-driven, and if future returns fall short of expectations, it could trigger a chain reaction. He made it clear that this deal has not alleviated previous concerns, but has instead made the problem more complex. In his view, the current "boom" is built on highly uncertain future earnings, and the accumulation of off-balance-sheet liabilities has reached an alarming level.