Moody's warns AI capex boom eroding credit quality at hyperscalers, projects $1T spending by 2027

Moody's Ratings warned that the trillion-dollar-a-year AI buildout is forcing hyperscalers like Alphabet, Microsoft, Amazon, and Meta to lean heavily on debt, stock sales, and off-balance-sheet financing to fund infrastructure. The agency projects capex will hit $785 billion in 2026 before reaching roughly $1 trillion in 2027, marking an unprecedented shift from the asset-light software model that built these companies' fortress balance sheets.

Direct debt across six tracked hyperscalers has reached approximately $460 billion, with long-term data center lease commitments ballooning to $1.2 trillion—more than $820 billion from leases not yet started. Moody's considers these obligations as debt-equivalent liabilities that will bind companies to significant future rent payments, even though they don't show up on balance sheets as traditional debt.

Moody's flagged structural circularity in the AI boom: hyperscalers have invested billions into AI labs (OpenAI, Anthropic) that spend heavily on cloud computing from those same companies, creating overlapping customer dependencies on the same demand assumptions. The agency also noted that Oracle (Baa2 with negative outlook) and specialized cloud provider CoreWeave are under immediate pressure, sitting closer to junk status.

Despite the warnings, Moody's said Microsoft, Alphabet, Amazon, and Meta retain among the world's strongest corporate balance sheets, making rating downgrades unlikely in the near term. However, investors will increasingly focus on these companies' ability to realize adequate returns on their massive capital commitments—a structural financial pivot unlike anything in the cloud era.