Moody's: Trillion-dollar AI shadow debt warning
The dizzying race for Artificial Intelligence (AI) dominance has created a financial blind spot that could compromise the stability of the world's largest technology corporations. According to a Moody's Ratings report, giants like Amazon, Meta, Alphabet (Google), Microsoft, and Oracle have accumulated data center lease commitments worth close to a trillion dollars. The alarming finding is that more than two-thirds of that figure (USD 662 billion) does not appear on their current balance sheets, because it corresponds to infrastructure that has not yet been built or has not begun commercial operations.
This "off-balance sheet" financial structure allows companies to keep these obligations hidden from investors and regulators under US GAAP accounting standards. According to Moody's analysts, this invisible debt is equivalent to 113% of the combined adjusted debt of these five firms at the end of 2025. By using residual value guarantees and contracts with short initial terms but unregistered "probable" renewals, companies manage to present financial statements that are stronger than their operational reality suggests, masking the real risk of an infrastructure supercycle that consumes capital at an unprecedented speed.
The economic risk is tangible due to the asymmetry between the useful life of technology and lease terms. While AI hardware, such as chips and servers, becomes obsolete within 4 to 6 years, data center lease contracts typically extend for 10 or 15 years. This disparity creates critical uncertainty: if the AI market cools down or if technology evolves towards models less dependent on large physical centers, these companies will be forced to pay millions in guarantees for depreciated assets. Moody's already plans to adjust its credit ratings to reflect these probable cash outflows, a move that could alter the cost of capital for the technology sector as a whole.
To gauge the scale of the problem, the total projected spending on these data centers is comparable to the Gross Domestic Product (GDP) of nations like Sweden or Singapore. As Alphabet reports massive quarterly jumps in its uninitiated commitments and Meta projects obligations extending until the end of the decade, the sector faces an accounting time bomb. Transparency in these commitments will be vital to prevent the rise of artificial intelligence from ultimately causing a solvency crisis in the companies that today lead the global digital economy.
Tags: Artificial Intelligence, Moody, AI infrastructure, Amazon, Meta, Alphabet, Microsoft, Oracle