Moody's Ratings revealed that the five leading hyperscalers, Amazon.com Inc. (NASDAQ:AMZN), Meta Platforms Inc. (NASDAQ: META), Alphabet Inc. (NASDAQ:GOOG) (NASDAQ:GOOGL), Microsoft Corp. (NASDAQ:MSFT) and Oracle Corp. (NYSE:ORCL), have racked up $662 billion in future data center lease obligations that remain off their balance sheets.
The race to build AI infrastructure has led to a significant financial overhang. These tech giants have entered into long-term data center lease commitments that have yet to commence and, therefore, are not classified as current liabilities. As a result, the obligations do not appear on their balance sheets under generally accepted accounting principles (GAAP), Fortune reported, citing a Moody's report released Monday.
By 2025 end, the companies had accumulated $969 billion in total undiscounted future lease commitments, highlighted Moody’s. In the coming years, as these leases take effect, more than half a trillion dollars in data center assets will begin appearing on corporate balance sheets, putting pressure on traditional accounting metrics.
Moody's analysts David Gonzales and Alastair Drake said the unrecorded $662 billion amounts to roughly 113% of the combined adjusted debt of the five hyperscalers. Gonzales told Fortune that the companies are not sidestepping a liability; rather, the obligation has not yet been recognized because the related services have not been delivered.
According to the report, the shift is driven by AI hardware's shorter lifespan. While U.S. data center leases traditionally ran 10–15 years, AI equipment typically lasts only four to six years, prompting hyperscalers to seek shorter initial lease terms with renewal options. To make these deals viable for landlords, they're often supported by substantial off-balance-sheet guarantees from the tenant.
Amazon, Microsoft, Meta, and Google did not immediately respond to Benzinga‘s requests for comment.
AI Data Center Accounting Concerns
The rapid growth of AI infrastructure in a short span has raised several questions. Fernando De Leon, founder of Leon Capital Group, has expressed concerns over the financial structure supporting this expansion. De Leon questioned why the world's largest companies are reluctant to own data centers outright. He argued that, despite AI being central to their business, major hyperscalers prefer to have others build and finance the assets instead of keeping them on their balance sheets.
Meanwhile, Michael Burry, the famed "Big Short" investor, has warned that tech giants like Microsoft, Meta, and Alphabet are using aggressive accounting maneuvers to mask the true costs of the AI infrastructure race. He claims this is boosting their profits by roughly 20%, accusing the companies of "understating depreciation" to shield their bottom lines.
The core allegation is that some companies are overstating how long their AI hardware remains useful. Despite rapid technological change, the claim is that they are extending the reported "useful life" of chips and servers to delay expenses and avoid impacting current earnings.
Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.
Image via Shutterstock
