The New AI Credit Playbook

Summary

Vishy Tirupattur discusses the rapid expansion of AI-driven CapEx and how credit markets are adapting to finance it. He highlights the dramatic upward revision in hyperscaler spending, broader and more innovative credit channels, and emerging physical and political constraints that will shape the buildout. The thesis is that credit markets will play a defining role in funding AI infrastructure for years to come.

• Hyperscaler CapEx projections have surged from $450B to $800B (2026) and $1.2T (2027) within a year.
• Over $200B in public AI-related credit issuance occurred in the first five months of the year.
• Issuance has expanded beyond USD into euros, sterling, Swiss francs, yen, and Canadian dollars.
• Issuer base now includes data center REITs and neo clouds in high yield markets.
• GPU financing is migrating from equity to credit markets via syndicated loans and ABS structures.
• Structural innovations combine project finance, tranching, residual value guarantees, and hyperscale-backed high yield debt.
• Grid access, power generation equipment, labor, and permitting delays are becoming significant bottlenecks.
• Energy infrastructure financing is expected to become more tightly linked to AI infrastructure financing.