Moody's Warns AI Spending Threatens Big Tech Credit Quality
Moody's Ratings has cautioned that "unprecedented" spending on artificial intelligence infrastructure is eroding free cash flow and increasing balance sheet risk for major technology firms.

Moody's Ratings has issued a warning that the significant and "unprecedented" buildout of artificial intelligence infrastructure, projected to reach an annual $1 trillion, is diminishing free cash flow and heightening balance sheet risks for major technology companies. The ratings firm noted in a research report that these expenditures are compelling even the world's largest tech firms, including Alphabet and Microsoft, to rely on debt, stock sales, and off-balance-sheet arrangements to fund their AI ambitions.
The shift from previously asset-light structures to a more capital-intensive model requires substantial investment, according to the report. Moody's is closely monitoring six entities: Microsoft, Amazon, Alphabet, Meta, Oracle, and CoreWeave, whose credit quality is considered to be under threat from these financial maneuvers, despite the companies maintaining strong financial positions.
Moody's forecasts that capital expenditures, covering investments in physical assets such as data centers, will total $785 billion in 2026 and approach $1 trillion the following year. This marks a departure from the long-standing Silicon Valley model, where software yielded high profit margins and robust balance sheets. Generative AI, conversely, necessitates a considerable physical footprint, including substantial investment in expensive and energy-intensive servers and chips.
To finance this accelerated expansion, tech corporations are increasingly turning to debt and public markets. Moody's reports that direct debt across the six monitored hyperscalers has reached approximately $460 billion. Companies like Alphabet have also undertaken significant equity sales to fund their AI infrastructure development.
Moody's highlighted that the free cash flow within the sector is facing pressure due to the massive upfront investments required for AI hardware and infrastructure, coupled with a longer revenue realization timeline. Lease commitments for data centers have also grown substantially, totaling an estimated $1.2 trillion, with over $820 billion from leases yet to commence. While these are not recorded as traditional debt, Moody's views them as debt-equivalent liabilities.