Five major tech giants—Alphabet, Microsoft, Amazon, Meta, and Oracle—are hiding $1.65 trillion in off-balance-sheet AI debt. According to a recent Nikkei study [nikkei.com], this unlisted financial pile now eclipses the $1.35 trillion they officially report.
To fund the AI data center boom, these tech giants package immense debt for chips, power, and servers into separate legal structures like joint ventures. For instance, Meta's off-balance-sheet obligations [thenextweb.com] sit at roughly $420 billion, nearly triple its reported debt. Meanwhile, Oracle holds $260 billion in future lease commitments, and Nvidia carries $119 billion in unlisted purchase obligations.
While this accounting treatment is entirely legal, market analysts at Morgan Stanley [businessinsider.com] have flagged massive data center lease expansions as a major credit risk factor. They note that global AI-related debt issuance is projected to hit $570 billion, making credit risk severely undervalued in the market. This rapid buildout introduces systemic risk; if AI demand falls short of exponential growth expectations, data centers will face steep asset markdowns. Furthermore, short 18-to-36-month AI server replacement cycles clashing against long-term 5-to-20-year bond durations create a massive refinancing cliff.
Rating agencies are already reacting to these stretched financial burdens. S&P has already downgraded Oracle's credit rating, while Moody's has also issued warnings regarding the expanding weight of pre-operational leases. As tech earnings drop over the next two weeks, the official numbers will look tidy, but investors are seeing less than half of the true industry leverage.