Artificial intelligence has become the biggest spending race the technology sector has ever seen. Microsoft (NASDAQ:MSFT | MSFT Price Prediction), Meta Platforms (NASDAQ:META), Amazon (NASDAQ:AMZN), Alphabet (NASDAQ:GOOG), and Oracle (NASDAQ:ORCL) are committing hundreds of billions of dollars to new data centers, networking equipment, and power infrastructure as they compete for AI leadership. Investors have largely embraced those investments because revenue continues to grow alongside demand for AI services.
Yet a recent investigation by Nikkei shows much of the financial commitment behind that expansion isn’t sitting where most investors expect to find it. As four of those five companies prepare to report quarterly earnings over the next week, shareholders may be evaluating balance sheets that reveal less than half of the financial picture.
The Debt You Won’t Find on the Balance Sheet
The numbers reported each quarter remain accurate. They’re also incomplete without reading the footnotes.
According to Nikkei’s review of SEC filings, the five technology giants have accumulated roughly $1.65 trillion in future lease and purchase obligations tied primarily to AI infrastructure. These commitments are perfectly legal under U.S. accounting rules because they represent future contractual obligations rather than traditional borrowings. Still, they don’t receive the same attention as reported debt.
The differences are striking:
| Company | Reported Debt | Est. Off-Balance Sheet Debt |
| Meta Platforms | $140 billion | ~$420 billion |
| Oracle | ~$100 billion | ~$273 billion |
| Microsoft | ~$100 billion | ~$350 billion |
| Amazon | ~$180 billion | ~$350 billion |
| Alphabet | ~$30 billion | ~$250 billion |
Meta’s obligations are roughly three times its reported debt. Oracle’s off-balance-sheet commitments have expanded about 30-fold in just four years as it races to build AI capacity.
None of this violates accounting standards. The concern is whether investors focusing only on reported debt are underestimating the financial commitments already made.
AI Optimism Makes the Numbers Look Comfortable
Many of these contracts finance data centers through long-term leases, project financing, and private credit rather than traditional corporate borrowing. That structure spreads financing across developers, insurers, and institutional lenders while allowing technology companies to avoid loading every obligation directly onto today’s balance sheet.
Ironically, that can make leverage appear lower precisely when spending is reaching record levels.
This matters because Meta, Microsoft, Alphabet, and Amazon all report earnings within days. Their reported debt ratios may appear manageable, while hundreds of billions of dollars in future payment obligations remain buried in the notes accompanying their SEC filings.
That doesn’t mean investors should panic. These companies also generate enormous cash flows. Microsoft produced nearly $100 billion in operating cash flow over the past year, while Alphabet, Meta, and Amazon each generated tens of billions of dollars that help support these long-term commitments.
The bigger issue is transparency rather than solvency.
The Risk Depends on AI Demand Staying Strong
The accounting changes only when those facilities begin operating. At that point, lease obligations move onto financial statements, depreciation begins, and any underutilized facilities can become impairment charges if demand falls below expectations. In other words, today’s hidden commitments become tomorrow’s reported assets and liabilities.
If AI adoption continues expanding at its current pace, those investments could generate attractive returns and justify every dollar committed. Granted, that’s exactly what management teams are betting on.
If enterprise AI spending disappoints, however, some data centers may never earn the returns originally projected. Any write-downs would ultimately affect shareholders, while lenders, insurers, and private-credit investors that financed the construction would also absorb losses.
Key Takeaway
In short, the headline isn’t that Big Tech has discovered a way to hide debt illegally — it hasn’t. These off-balance-sheet commitments are disclosed in SEC filings and comply with accounting rules. The real takeaway is that investors who rely only on headline debt figures are missing a large portion of the AI spending story.
As earnings season unfolds, reported debt will probably look comfortable. The $1.65 trillion of contractual commitments likely won’t dominate earnings headlines. Smart investors should look beyond the income statement and balance sheet into the footnotes. That’s where the full scale of Big Tech’s AI bet — and the risks that come with it — is waiting to be found.
Contact [email protected] for any questions or corrections.
Top Gaining Stocks
Top Losing Stocks