US hyperscalers have issued approximately $220 billion in debt linked to AI spending as of mid-August 2026, compared with $12.5 billion in the same period of 2025, a roughly 17-fold increase that is testing demand in the investment-grade bond market. Amazon recently priced a $25 billion long-dated offering at about 120 basis points over Treasuries, wider than in prior years, while tech investment-grade spreads stand at 89 basis points—nine basis points above the broader market. New deals have required 10 to 15 basis points in concessions amid late-2026 issuance fatigue cited by DWS’s George Catrambone. Goldman Sachs has estimated AI-related debt could reach about $489 billion by mid-2026, and Morgan Stanley’s global forecast runs as high as $570 billion. The four largest U.S. tech firms had already issued more than $170 billion earlier in the cycle, Broadcom has sought nearly $100 billion for AI chip and infrastructure financing, and record investment-grade supply near $1.7 trillion is colliding with government funding needs, higher long-term Treasury yields and a rising cost of capital that could pressure secondary tech bond prices if concessions widen further.