Big Tech’s Hidden AI Debt: $1.65tn Unaccounted for
Five tech giants—Alphabet, Microsoft, Amazon, Meta, and Oracle—are carrying a staggering $1.65 trillion in off-balance-sheet debt, more than their reported figures show. This unseen debt is funding the AI data center boom through a legal yet questionable accounting trick that Enron utilized before its collapse 25 years ago.
The Enron Echo
While the amounts reported by these companies appear manageable, a Nikkei study reveals that hidden debt totals approximately $1.65 trillion—a significant increase from the $1.35 trillion they officially report. The same structures used to hide debt during Enron’s era are being employed today, albeit within legal boundaries due to tighter regulations and increased disclosure requirements.
How It Works
The process involves creating separate legal entities, often joint ventures, to package debt for chips, servers, and power. Meta’s Hyperion data center in Louisiana is a prime example. Meta and Blue Owl Capital contributed equity to a structure that took on $27 billion in debt, with Meta as the sole tenant. They argue that they don’t have to record this debt because they are not responsible for finding replacement tenants.
Oracle has $260 billion in future lease commitments that will eventually appear on its books, while Nvidia carries $119 billion in purchase obligations. Alphabet and Microsoft also maintain their vehicles off-book.
The Numbers Behind the Boom
The scale is remarkable: Meta’s off-balance-sheet debt alone totals about $420 billion—triple its reported debt. Oracle’s has grown thirtyfold in four years. This massive investment is driven by a race to build and equip AI data centers, expected to surpass $3 trillion through 2028.
Implication and Concerns
The timing is sensitive, as earnings reports for four of these giants are due within the next two weeks. The $1.65 trillion in footnotes will likely go unnoticed in the short term. However, the true test lies ahead: When data centers come online, their leases transition to the balance sheet, and if AI demand falls short, lenders and insurers bear the risk.
Some analysts already sound the alarm. S&P has lowered Oracle’s credit rating due to excessive leverage, while Morgan Stanley and Moody’s have flagged the broader issue. "What if one of these companies was a house of cards," questioned accounting consultant Tom Selling, "and was propping itself up with this accounting treatment?"