Hyperscaler off-balance sheet obligations hit $3 trillion as AI spending spree accelerates

Hyperscaler off-balance sheet obligations have reached $3 trillion, according to CNBC's Sara Eisen, as Microsoft, Amazon, Meta, Alphabet, and Oracle finance AI infrastructure through hidden debt. A Moody's analysis from February 2026 estimated $662 billion in unrecognized lease commitments, while a Nikkei Asia report from July 2026 pegged combined hidden obligations at $1.65 trillion, exceeding their disclosed on-balance-sheet debt of roughly $1.35 trillion. Global AI data center capital expenditure is projected to exceed $3 trillion over the next several years, with about half funded by debt and off-balance sheet vehicles.

Via datacenterknowledge.com Hyperscaler off-balance sheet obligations hit $3 trillion as AI spending spree accelerates The tech giants building the AI future are financing it with a mountain of hidden debt that dwarfs what shows up in their SEC filings CNBC’s Sara Eisen reported that hyperscaler off-balance sheet obligations have now reached $3 trillion, a figure that should make anyone holding big tech stocks pause and do some math. The number captures the staggering scale of financial commitments that Microsoft, Amazon, Meta, Alphabet, and Oracle have quietly accumulated as they race to build the AI infrastructure of tomorrow. The hidden debt machine powering AI A Moody’s analysis from February 2026 estimated that hyperscalers collectively held approximately $662 billion in lease commitments that had yet to be recognized on their balance sheets. Total undiscounted future lease commitments reached $969 billion according to the same analysis. A separate Nikkei Asia report from July 2026 pegged the combined hidden obligations of five major tech giants at $1.65 trillion. That figure, which includes both leases and purchase commitments, exceeded their disclosed on-balance-sheet debt of roughly $1.35 trillion. The debt you can’t easily see is larger than the debt you can. Microsoft announced a roughly $100 billion AI Infrastructure Partnership structured as a special purpose vehicle in 2024, utilizing dedicated debt facilities. Where the money comes from Global AI data center capital expenditure is projected to exceed $3 trillion over the next several years. According to the analysis, roughly half of that funding is expected to come from operating cash flow. The other half will be sourced through debt instruments and off-balance sheet vehicles. Moody’s has already flagged potential risks tied to credit quality and debt transparency. If these obligations start migrating onto balance sheets, either because leases commence or accounting standards tighten, the financial profiles of these companies could look meaningfully different overnight. What this means for investors If $1.65 trillion in hidden obligations exceeds $1.35 trillion in disclosed debt, the true leverage of these companies is roughly double what a casual glance at their balance sheets would suggest. Credit rating actions are the canary in this particular coal mine. If Moody’s or its peers start adjusting their treatment of off-balance sheet obligations for hyperscalers, it could trigger a cascade of re-evaluations across the sector. Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy https://cryptobriefing.com/editorial-policy/ .