Morgan Stanley projects AI-linked bond issuance will reach nearly $570 billion in 2026, with $236 billion already placed by May 31 — four times the 2025 pace [1]. The forecast transforms the AI infrastructure boom from a capital expenditure story into a credit story, with consequences that extend beyond the technology sector.

Hyperscaler capital expenditure now runs close to 100 percent of operating cash flows, versus a ten-year average of 40 percent [1]. Companies like Amazon, Alphabet, Meta, Microsoft, and Oracle are issuing bonds to fill the gap that free cash flow can no longer cover. Amazon set records with a C$14 billion Canadian maple bond and the largest euro corporate bond ever at 14.5 billion euros.

The structural shift is systemic. AI-linked debt surpassed US banks as the largest investment-grade market segment by October 2025 [1]. The $4.8 trillion in target-date fund assets that track bond indexes must proportionally absorb every new AI bond, meaning a sector-specific credit event would propagate into retirement accounts without any individual investor choosing the exposure.

S&P has warned that Amazon will likely post negative free operating cash flow over the next two years while carrying $364 billion in cloud backlog and $200 billion in planned 2026 capital spending [1]. May 2026 inflation sits at 4.2 percent, the highest since April 2023.

The question the paper poses is whether AI revenue can cover the interest. At $570 billion in bonds, the annual debt service cost is substantial. If AI revenues disappoint, the refinancing cycle becomes a crisis.

-- THEO KAPLAN, San Francisco