Oct 9, 2026 · 45 min
AI infrastructure debt exposes a fragile growth story
“Banks Are Herd Animals” | Ed Zitron on the AI Debt Slow Boil
The episode examines whether borrowing, guarantees, and optimistic revenue claims are sustaining AI investment beyond what durable demand can support.
- 1Rising Treasury yields are increasing financing costs for data centers, chips, and other capital-intensive AI projects.
- 2Oracle’s financing structures and force majeure notice raise questions about hidden leverage and the viability of planned infrastructure.
- 3Hyperscaler guarantees and annualized revenue claims may be encouraging lenders to underprice uncertain demand and project risk.
Don't miss
Ed Zitron explains why Oracle’s force majeure notice for Project Jupiter is unusual for a data center and potentially signals deeper financing stress.
The brief
Ed Zitron and Will Guy examine why the AI boom may deflate gradually, as enormous compute commitments and planned data-center capacity meet uncertain demand.
Higher Treasury yields raise borrowing costs across the economy, making AI infrastructure harder to finance even when companies do not directly buy government bonds.
Oracle’s force majeure notice for Project Jupiter becomes a warning sign: an unusual construction-related measure that may point beyond an ordinary delay.
The discussion argues that guarantees from hyperscalers can make risky projects appear safer, encouraging lenders to follow the crowd rather than test future demand.
Oracle’s chip-leasing structure and optimistic revenue run rates sharpen the central concern: accounting and projections may obscure the economics beneath AI’s spending surge.