Focus: Banks get creative and look further afield as AI-fueled debt soars

• Hyperscaler capital expenditures are estimated at $725 billion in 2026, nearly double mid-2025 levels, growing faster than operating cash flow and requiring external funding
• Amazon raised a record €14.5 billion ($16.56 billion) in euros in March, while Alphabet set borrowing records across yen, sterling, and Swiss franc markets
• Data center lease-backed deals have emerged as a new financing structure, with approximately 15 such deals sold since last year, including an $810 million note by Stingray Compute that was nine times oversubscribed

Market Summary: AI-Fueled Corporate Debt Surge Drives Banking Innovation

Key Developments

AI-related debt has reached nearly 15% of investment-grade bond issuance in 2026, prompting banks to develop creative financing solutions as technology companies fund massive infrastructure buildouts. Hyperscalers including Amazon and Alphabet have issued $60 billion in bonds across multiple currencies over the past 12 months.

Major Transactions and Figures

Amazon set a European record with a €14.5 billion ($16.56 billion) eight-part euro bond deal in March, the largest ever in that market. Alphabet broke borrowing records across yen, Swiss franc, and sterling markets, while also completing the largest Canadian dollar deal from a tech company since 1997.

Capital expenditures for hyperscalers are projected at approximately $725 billion in 2026, nearly double mid-2025 levels, according to BNP Paribas. This spending is outpacing operating cash flow, driving external funding needs.

Innovative Financing Structures

Banks are structuring deals around pre-arranged data center leases to provide cash flow visibility. Stingray Compute recently issued $810 million in notes backed by an Amazon data center lease, with the offering nine times oversubscribed. Approximately 15 such construction loan-inspired deals have been sold to high-yield investors since last year.

Market Implications

Investment-grade issuance could exceed $2 trillion for the first time in 2026, according to Morgan Stanley. While demand remains strong for high-quality hyperscaler bonds, concerns are emerging about market saturation as companies repeatedly tap debt markets.

However, analysts note AI-related debt remains low relative to broader investment-grade credit indices. Portfolio managers emphasize that as long as hyperscalers continue long-term AI investments, funding demand will persist.