Moody's Ratings has put the market on notice: the trillion-dollar-a-year race to build artificial intelligence infrastructure is cracking the financial foundations of the very companies meant to benefit from it. In a research note this week, the firm said the spending surge is forcing even the world's most cash-rich corporations to lean heavily on debt, equity sales and off-balance-sheet maneuvers to fund their AI ambitions, a shift that "threatens credit quality" across the six hyperscalers it tracks, Microsoft, Amazon, Alphabet, Meta, Oracle and CoreWeave.
The numbers are staggering by any standard, let alone for an industry that spent decades perfecting the asset-light model. Moody's projects capital expenditures will hit $785 billion this year before crossing the $1 trillion threshold in 2027. Direct debt across the group has already reached roughly $460 billion. Alphabet alone tapped public markets for an $85 billion equity sale last month. The transition from software margins to server farms is not a pivot; it is a balance-sheet event.
What does not show up as debt looks a lot like debt. Lease commitments across the six companies have ballooned to $1.2 trillion, with more than $820 billion tied to data centers still under construction. Moody's treats these as debt-equivalent liabilities, binding the hyperscalers to rent payments long after the GPUs inside have depreciated. The asset-light playbook, high margins, fortress balance sheets, minimal capex, has been inverted. Generative AI demands a vast physical footprint: warehouses crammed with expensive, energy-hungry chips that must be paid for up front while revenue materializes over a longer, uncertain horizon.
The immediate pressure is not evenly distributed. Microsoft, Alphabet, Amazon and Meta retain among the strongest corporate balance sheets in the world, making imminent rating downgrades unlikely. The strain is concentrated at the lower end: Oracle sits at Baa2 with a negative outlook, just two notches above junk, while CoreWeave operates in high-yield territory at Ba3, financing its GPU fleets through complex private debt structures. The market is already pricing the difference.
Moody's also flagged a structural circularity that should give investors pause. Multibillion-dollar backlogs at the hyperscalers stem partly from strategic deals with pre-IPO AI labs such as OpenAI and Anthropic, labs the hyperscalers themselves have funded with billions. Those labs then spend heavily on cloud capacity from their backers, creating a self-referential loop where revenue and capex are effectively recycled within the same ecosystem. The overlapping relationships heighten risk because the industry's biggest companies are increasingly dependent on the same AI customers and the same assumptions about future demand.
The irony is that Wall Street is profiting handsomely from financing this transition, underwriting the debt and equity offerings that keep the buildout moving. But the ratings signal is clear: the credit metrics that once made these names untouchable are deteriorating by design. The question is not whether the hyperscalers can afford the spend, they can, but whether the returns on a trillion dollars of annual capex will ever justify the leverage, the lease obligations, and the circular revenue that currently props up the growth story.