Five Big Tech Giants Are Hiding $1.65 Trillion in Debt — AI's Enron-Style Accounting Trick
The AI arms race is burning $3+ trillion — and behind it lies an even deeper ledger.
On July 22, Nikkei Asia published an investigation revealing a hidden ledger beneath AI's glossy surface: five US tech giants — Alphabet, Microsoft, Amazon, Meta, and Oracle — are concealing an estimated $1.65 trillion in off-balance-sheet debt. That figure is 122% of the $1.35 trillion they officially report on their balance sheets. In other words, these five companies owe far more than they admit.
The key mechanism is the variable interest entity (VIE) — a special-purpose vehicle that allows companies to keep data center construction loans, equipment leases, and energy contracts off the parent company's financial statements. Meta alone formed a VIE joint venture last year to build a Louisiana data center, with a maximum exposure of $46 billion — and recently expanded the project toward $250 billion. Alphabet keeps data center leases and credit backstops off its books through similar VIEs. Microsoft provides few details about its VIEs, stating only that it does not consolidate those entities.
Anyone familiar with accounting history will recognize the pattern. In 2001, energy giant Enron collapsed under the weight of hundreds of billions in hidden debt parked in nearly identical off-balance-sheet structures — one of the largest corporate scandals in US history. Now the world's most valuable companies are using the same playbook to finance their AI spending spree. "The accounting treatment itself is in fashion," Tom Selling, a technical accounting consultant, told Bloomberg Tax. "But what if one of these companies was a house of cards and was propping itself up with this accounting treatment? To me, that's the risk."
Nikkei's investigation found that the five giants' hidden debt has ballooned eightfold in four years — from roughly $200 billion in 2022 to $1.65 trillion today. Behind it all: the AI industry's plan to spend over $3 trillion on infrastructure. Analysts warn that if AI demand falls short of expectations, these off-the-books obligations could trigger a chain reaction.
Sources: Nikkei Asia, Bloomberg Tax, Futurism, Tom's Hardware
BenPig tried to count — 1.65 trillion is a 1 with 12 zeros.