Hyperscaler debt shows mounting risks as market reprices AI companies
Private equity firm Apollo Global Management warned Wednesday that corporate debt issued by major cloud computing companies powering the AI boom is becoming riskier. Credit default swaps, or insurance contracts against bond defaults, have become more expensive for bonds issued by hyperscalers like Amazon, Meta, Google and Microsoft. Apollo chief economist Torsten Slok says this repricing reflects weakening credit fundamentals including debt-financed AI capital spending, rising leverage, negative free cash flow and uncertain returns on infrastructure investments. The gap between hyperscaler and bank credit default swaps has widened to about 60 basis points since October 2025, indicating independent risk increases rather than routine dealer hedging. While some technology investors argue margins are improving and it is too early to worry, economists including Dean Baker note that sophisticated credit investors are pricing in substantial risk that major AI companies may struggle to make good on their debt commitments.
Key facts
- Apollo Global Management warned September 16, 2026 that credit default swaps on hyperscaler debt show rising risk levels
- CDS spreads between hyperscalers and banks widened 60 basis points since October 2025, up from near zero
- Amazon has 23 percent debt-to-equity and negative free cash flow of 30 billion dollars
- Meta has 34 percent debt-to-equity and negative free cash flow of 25.7 billion dollars
- Microsoft has 7.34 percent debt-to-equity and positive free cash flow of 33.4 billion dollars
- Hyperscaler bond issuance reached 244 billion dollars through July 2026, more than double 108 billion in all of 2025
- Five major US tech companies have accumulated approximately 1.65 trillion dollars in off-balance-sheet debt obligations
- Chief economist Torsten Slok cited debt-financed AI capex cycle, rising leverage, negative cash flow and uncertain asset payback as repricing factors
How different outlets framed it
Left-leaning coverage: Progressive economists and analysts focus on structural vulnerabilities in AI financing. They point to hidden off-balance-sheet debt exceeding 1.6 trillion dollars and binding commitments for GPU purchases that don't appear as traditional debt. Concerns center on whether tech companies can realistically monetize investments, questioning whether this mirrors the 2008 housing crisis dynamics with debt-funded assets generating uncertain returns.
Right-leaning coverage: Conservative observers emphasize the competitive dynamics draining capital from safer investments like Treasuries. They highlight how massive hyperscaler debt issuance is pushing government borrowing costs higher, affecting mortgage rates and business financing. Some cite this as evidence of tech sector overreach and question whether the federal government should be concerned about capital crowding effects, though hyperscalers have so far maintained access to debt markets.
Limits of this briefing
Sources emphasize that not all hyperscalers face equal risk, with Microsoft notably maintaining positive cash flow while others struggle. Debate persists over whether this represents a genuine credit crisis or temporary market repricing ahead of AI revenue materialization. Off-balance-sheet obligations and shadow lending structures make full risk assessment difficult.
Sources
- CNBC: Hyperscaler debt signals warning sign, Apollo cautions Centre
- Investing.com: Apollo warns cloud computing debt grows riskier amid AI spending Centre
- The Hill: Tech companies rack up debt, risks to fund AI ambitions Centre
- Fortune: Wall Street used to worry about crowding out, but hyperscalers are reversing it Centre
- BTC.com: Economist compares hyperscaler AI debt stress to 2008 housing collapse Left