Big Tech's Off-Balance-Sheet AI Debt Hits $1.65 Trillion

Five major U.S. technology companies have accumulated approximately $1.65 trillion in off-balance-sheet debt tied to artificial intelligence infrastructure investments, according to a study by Nikkei. The hidden liabilities have grown eightfold over roughly four years and now exceed the companies' transparent, reported debt.

The opaque financing arrangements stem primarily from data center leases and graphics processing unit supply contracts required to support AI operations. At Meta, off-balance-sheet obligations total about $420 billion—nearly triple the company's on-balance-sheet debt, Nikkei reported.

Why it matters

Investors and analysts typically rely on balance sheets to assess corporate financial health and risk exposure. When companies structure AI investments through operating leases and long-term supply agreements rather than direct purchases, those commitments don't appear as traditional debt. The scale of these hidden obligations—$1.65 trillion across just five firms—suggests the true capital intensity of the AI race may be significantly underestimated in public markets. This opacity complicates risk assessment at a time when AI spending continues to accelerate across the technology sector.

The scale of hidden commitments

The $1.65 trillion figure represents obligations that don't appear on corporate balance sheets under current accounting standards. These arrangements include:

• Long-term data center lease agreements
• Multi-year GPU supply contracts with manufacturers
• Infrastructure commitments tied to AI model training and deployment

Oracle is among the other companies identified in the Nikkei analysis as carrying substantial off-balance-sheet AI-related debt, though specific figures for the other three companies were not detailed in the report.

Implications for transparency

The rapid accumulation of these hidden liabilities raises questions about financial disclosure practices in the technology sector. Traditional debt metrics may no longer provide an accurate picture of companies' true financial obligations when massive infrastructure commitments are structured to remain off the balance sheet.

As AI investments continue to grow—with some industry leaders projecting annual AI spending could reach $5 trillion—the gap between reported and actual financial commitments may widen further. This creates challenges for investors attempting to compare companies or assess the sustainability of current AI investment levels.

The findings were first reported by Nikkei, based in Palo Alto, California.