Record AI-debt issuance is overwhelming the market. Global AI-linked bond issuance reached ~$236 billion by the end of May 2026, and Morgan Stanley projects $570 billion by year-end — a fourfold year-over-year increase . Hyperscalers are the dominant borrowers
.
Credit spreads are under pressure. Despite overall IG spreads still being tight, four of the five largest hyperscalers have underperformed the broader market on a spread basis year-to-date . Free cash flow compression and rising leverage are creating "nearterm pressure on spreads," with bond investors increasingly questioning the return on AI capital
.
Investor pushback is intensifying. Funds are demanding higher premiums even from AAA-rated tech issuers, and the sheer supply of hyperscaler debt is "potentially stalling the overall grind tighter in credit spreads" .
On July 22, 2026, Alphabet reported a record Q2:
Yet the market's reaction was a simultaneous stock and bond sell-off. The capex hike to as much as $205 billion — alongside continued delays to Alphabet's flagship Gemini AI model — aggravated fears that spending is racing ahead of monetization . The $15 billion capex increase signaled to bondholders that Alphabet will keep flooding the market with debt, pressuring secondary prices on outstanding issues like the century bond
.
Bottom line: The 100-year bond's more-than-10% loss is a canary in the coal mine for the broader AI-debt cycle. Its extreme duration magnified rate moves, but the deeper story is that bond investors — who were exuberant in February — are now pricing in real doubt about whether the hyperscalers' unprecedented debt-funded capex will deliver the long-term returns needed to service that debt. Even a record cloud quarter and $119.8 billion in revenue could not arrest that anxiety.