Apollo Warns: Rising Credit Risk for Cloud Giants as CDS Spreads Widen to 60 Basis Points
nashnova research
Apollo's chief economist warns that hyperscaler CDS spreads have widened to 60 basis points, as debt-fueled AI capex systematically pushes up credit risk — and the equity market has yet to catch up with this repricing.
CDS spreads went from zero to 60 bps — what does that mean?
A CDS — credit default swap, essentially default insurance on a bond — spread measures how much the market charges to insure against a company's default. The wider the spread, the more worried the market is.
Since October 2025, the CDS spread between hyperscalers and banks has widened from near zero to roughly 60 basis points. This means → the bond market now sees these tech giants as meaningfully riskier than banks; six months ago, the two were priced as nearly equal.
Over the same period, hyperscaler stock prices did not fall in step — credit and equity have been persistently decoupling. In plain terms = bond investors are already voting "no," but equity investors haven't reacted yet.
Where is all the money going?
To lock in AI infrastructure, the four major hyperscalers are borrowing heavily to build data centers and compute capacity — capex far exceeds operating cash flow, and free cash flow has broadly turned negative.
FactSet data: Google free cash flow negative $25.7 billion, debt-to-equity 13%; Amazon negative $30 billion, debt-to-equity 23%; Meta negative $25.7 billion, debt-to-equity 34%.
The sole exception is Microsoft: free cash flow positive $33.4 billion, debt-to-equity just 7.34%. This reflects either more disciplined AI spending or stronger cash generation from its existing business — likely both.
Could the spread widening be just a technical artifact?
Apollo chief economist Torsten Slok explicitly ruled out the "new-issuance hedging" explanation. In plain terms = some might argue spreads widened simply because these companies issued a wave of new bonds and underwriters hedged — but Slok says that is not what happened.
His logic: if hedging were the cause, bank CDS spreads should have widened in tandem, because banks remain the single largest source of investment-grade bond supply. In fact, bank spreads have held steady at around 40 basis points, barely moving.
This means → the rise in hyperscaler credit risk is a standalone event — the market is repricing their own fundamentals, not reacting to technical noise.
What to watch next?
Credit markets are generally viewed as a more sensitive early-warning indicator than equities — bond investors have already cast their vote.
Several leading frontier-model companies have recently signaled they want to slow product iteration on safety grounds. If followed through, this directly hits the revenue outlook for the cloud providers that host these models.
The key checkpoint: whether the equity market catches up with credit-market repricing. If stock prices stay put for much longer, at least one of the two markets is mispricing the risk.
市场有风险，内容仅供研究参考，不构成投资建议。