Michael Burry questions Oracle's AI infrastructure disclosure practices
In brief
- Burry published Substack analysis of Big 5 hyperscalers' AI infrastructure lease commitments
- Oracle disclosed $261 billion in off-balance-sheet data center lease obligations
- S&P downgraded Oracle to BBB- in July 2026 over high AI investment leverage
- Burry maintains short positions against Oracle amid valuation concerns
Oracle's Undisclosed Commitments
Oracle reported fiscal Q1 2027 earnings on September 10 with cloud infrastructure revenue of $7.39 billion, a 121% year-over-year increase. The company raised its full-year revenue guidance to at least $90 billion on the back of the results. Yet according to Burry's analysis, Oracle has disclosed $261 billion in off-balance-sheet data center lease commitments, many of which have not yet commenced.
The distinction matters. Operating leases don't appear on balance sheets the way traditional debt does. They create future cash obligations without triggering the same level of investor scrutiny. Burry argues this structure masks the true financial exposure embedded in the race to build out AI infrastructure.
Credit Rating Signals Trouble
S&P downgraded Oracle to BBB- in July 2026, citing high leverage from substantial AI investments. That rating sits one notch above junk status. Oracle's share price declined after the September 10 earnings report, suggesting the market may be recalibrating its view of the company's fundamentals.
Burry has maintained short positions against Oracle's stock, a direct bet that the market is overvaluing the company relative to its actual financial health. His analysis extends beyond Oracle. For the Big 5 hyperscalers collectively, Burry's analysis suggests there may be trillions in such commitments that aren't receiving adequate investor attention.
"Lease commitments, particularly those structured as operating leases that don't appear on the balance sheet, can represent enormous future cash obligations without triggering the same investor scrutiny as traditional debt." — Michael Burry
The implication is clear: investors may be pricing in AI upside without fully accounting for the infrastructure debt required to deliver it. When that debt eventually surfaces—whether through covenant breaches, credit downgrades, or forced restructurings—valuations could contract sharply. Burry's track record suggests his skepticism warrants attention.
