Apollo Warns AI Spending Is Increasing Credit Risk for Major Cloud Providers

The corporate debt of major cloud computing companies is becoming riskier as they ramp up spending on artificial intelligence infrastructure, according to a warning from Apollo Global Management. The investment firm noted that the market is beginning to reprice the credit fundamentals of these tech giants, often called 'hyperscalers,' due to their aggressive, debt-financed capital expenditures.

Market Signals Rising Risk

A key indicator of this trend is the rising cost of credit default swaps (CDS), which serve as insurance against bond defaults. According to Apollo's research, the gap between the CDS spreads for hyperscalers and those for major banks has widened significantly.

This spread has grown to approximately 60 basis points from nearly zero since October 2025. Apollo suggests this indicates that the market perceives an increase in the standalone credit risk of these tech firms, rather than it being a byproduct of dealer hedging activities for new bond sales.

AI Spending Strains Balance Sheets

In a note, Apollo Chief Economist Torsten Slok attributed the repricing to a "debt-financed AI capital expenditure cycle." Slok highlighted concerns over rising leverage, negative free cash flow, and the uncertain returns on investment for rapidly depreciating AI assets.

Ad

Data from FactSet illustrates the financial positions of several key players:

• Alphabet (GOOGL): A forward debt-to-equity ratio of 13% and a projected negative free cash flow of $25.7 billion.
• Amazon (AMZN): A forward debt-to-equity ratio of 23% and a projected negative free cash flow of $30 billion.
• Meta Platforms (META): A forward debt-to-equity ratio of 34% and a projected negative free cash flow of $25.7 billion.
• Microsoft (MSFT): A forward debt-to-equity ratio of 7.34% and a positive free cash flow of $33.4 billion.

Broader Industry Context

Apollo's analysis comes as leaders in the AI industry have publicly discussed slowing the pace of developing advanced models due to safety concerns. Such a slowdown could have significant financial repercussions for the cloud providers who supply the essential computing power, potentially extending the payback period for their massive infrastructure investments.