Big Tech's $3 Trillion AI Commitment Hides Off Balance Sheets

Seven major technology companies have committed roughly $3 trillion to AI infrastructure spending through off-balance-sheet obligations, according to new research from Morgan Stanley. That figure sits on top of an estimated $770 billion in debt and lease obligations already recorded on their balance sheets.

The companies analyzed include hyperscalers Google, Meta, Microsoft, Oracle, and Amazon, plus chipmakers Nvidia and Broadcom. Their combined financial commitments to AI development dwarf the already substantial spending figures that have dominated headlines.

Why it matters

This analysis exposes a layer of financial obligation that doesn't appear in standard corporate reporting, making the true scale and leverage of AI infrastructure investment harder to track. The commitments also create a cascading effect through the supply chain, as suppliers and data center developers borrow against these guaranteed future payments to build capacity before receiving actual funds.

Breaking down the commitments

Off-balance-sheet commitments represent future payment obligations that don't count as official debt on corporate financial statements. Morgan Stanley's breakdown shows hyperscalers have committed $1.1 trillion specifically for data center leases that haven't yet started. All seven companies combined have agreed to purchase $1.7 trillion worth of chips, memory, and networking equipment.

These purchase agreements have accelerated dramatically. Google's commitments alone reached $707 billion in the most recent quarter, up from $72.5 billion for all of 2025.

The leverage cascade

The structure creates significant downstream effects. Suppliers and data center developers use long-term commitments from investment-grade tech companies as collateral to secure their own borrowing. This allows infrastructure to be built before the tech giants make payments or recognize liabilities on their books.

Todd Castagno, head of global valuation, accounting and tax at Morgan Stanley who coauthored the analysis, compared the situation to "developing a car market without ever having seen the capabilities of car before." The fundamental question for investors is timing—when these massive investments will generate returns in what remains a nascent market.

Uncertainty ahead

Several variables complicate the picture. Each company pursues different spending strategies with varying risk profiles. The timeline for deploying that $3 trillion remains unclear, and these commitments—often structured as contracts—could theoretically be renegotiated.

Eventually, these obligations will migrate onto official balance sheets. Whether the AI investments driving them will have proven profitable by that point remains the central question for investors and analysts watching this unprecedented capital deployment.

These findings were first reported by Axios, based on Morgan Stanley's analysis of corporate filings.