The consensus is clear: 2026 issuance is running at roughly double the 2025 pace, and accelerating.
A July 2026 fund manager survey captured a dramatic shift in institutional concern. 48% of investors named AI hyperscaler capital spending as the most likely trigger of a systemic credit event — the largest single answer, far ahead of private credit at 34% and consumer credit at 5% .
This finding is corroborated by a separate Bank of America survey from May 2026, which found 34% of fund managers pointing to AI hyperscaler capex as the top systemic risk, double the share from April . Bloomberg independently reported that AI data-center borrowing is "quickly climbing Wall Street's list of potential credit threats"
.
The AI financing wave has moved beyond investment-grade names into higher-risk territory. In May 2026, H.I.G. Capital-backed PolarDC Group raised a record €800 million ($913 million) senior secured Nordic high-yield bond — the largest such deal in Nordic market history . Key terms:
The proceeds will refinance existing debt on Polar DC’s first data center in Drangedal, Norway, and complete construction of two additional facilities in 2026 and 2027 . The deal has already spawned imitators: Pure Data Centres Group Ltd. (backed by Oaktree Capital) subsequently sought to raise €1 billion in the same market
, and Prime Data Centers has sought $600 million.
AI-related borrowing has become a structural force in the U.S. corporate bond market. A Reuters report from June 29, 2026, states that AI-related debt is close to 15% of total investment-grade bond issuance in the U.S. this year . The IESE analysis estimates that AI-related debt now accounts for roughly 30% of net new investment-grade supply in the U.S. dollar market, reflecting a broader definition
.
This concentration has implications for duration and liquidity: most issuances are long-dated (five years and longer), locking in funding for multiyear AI infrastructure programs and substantially extending the average duration of the IG market . The Dallas Fed has noted that new duration supply from AI-related issuance could total as much as $360 billion in 10-year equivalents in 2026, or about one-eighth of the duration supply from U.S. Treasury issuance
.
The AirTrunk SYD3 loan serves as a canary in the coal mine for regional bank exposure. The deal is explicitly structured as "Asia-Pacific lenders run closer to AI exposure limits" . The borrowing cost dynamic is reflected in Polar DC's pricing — 600 bp over Euribor — which signals that lenders are demanding higher premiums as sector concentration builds.
A PIMCO analysis from May 2026 noted that hyperscaler capex is now expected to absorb 94% of cash flow from operations in both 2026 and 2027, versus just 40% in 2023 . Rising capital spending combined with falling free cash flow signals a structural shift toward leverage that is testing bank balance sheets across Asia Pacific.
The AI infrastructure debt wave is real, accelerating, and concentrating risk. By mid-2026, nearly $500 billion in AI-related debt has been issued, institutional investors rank it as the top systemic credit threat, and the market is expanding into higher-yield, higher-risk territory. The deals themselves — from Blackstone’s A$4.3 billion syndicated loan in Sydney to Polar DC’s record Nordic bond — illustrate a financing ecosystem that is both remarkably creative and increasingly fragile. Two specific claims (a Chicago Fed tail risk warning and Citi analysts on deal complexity/opacity) were not captured in available search results and remain unverified in this analysis.