Hidden debt at America's largest technology companies has surged eightfold to $1.65 trillion over four years, driven by massive artificial intelligence infrastructure investments, according to a new Nikkei study.

The off-balance-sheet liabilities now exceed the companies' transparent debt, making it harder for investors to assess financial risk as AI spending accelerates across the sector.

Meta leads with approximately $420 billion in hidden debt — nearly triple its reported obligations. The social media giant's undisclosed liabilities stem primarily from long-term data center lease agreements and graphics processing unit supply contracts needed to power its AI ambitions.

Oracle also features prominently in the analysis, though specific figures for the database company were not disclosed in the available excerpt.

Why This Matters for AI Investors

The hidden debt surge reflects the enormous capital requirements of AI infrastructure buildouts. Companies are committing to multi-year contracts for cloud capacity, specialized chips, and data center space without fully disclosing these future obligations to shareholders.

These off-balance-sheet commitments include:

  • Long-term GPU supply agreements with Nvidia and other chip makers
  • Data center lease contracts spanning multiple years
  • Cloud infrastructure capacity reservations
  • Power and cooling system commitments

The opacity makes it difficult for investors to understand the true financial exposure of tech companies racing to build AI capabilities. Traditional debt metrics may significantly understate actual leverage and future cash flow commitments.

The study examined five unnamed US technology giants, suggesting the $1.65 trillion figure represents a conservative estimate of sector-wide hidden AI-related obligations.

This financial engineering comes as companies face intense pressure to demonstrate AI progress while managing investor expectations about profitability timelines. The hidden nature of these commitments allows firms to pursue aggressive AI strategies without immediately alarming shareholders about debt levels.

The findings arrive as AI infrastructure costs continue climbing, with some estimates suggesting the sector requires trillions in annual investment to maintain competitive positioning in the global AI race.