Forward Guidance The AI Bubble Is Widely Misunderstood | Steve Hou
37 snips
Apr 29, 2026 
Steve Hou, a senior quant researcher at Bloomberg with a PhD in macro and financial econometrics, explains why AI feels as big as the internet. He breaks down the massive AI-driven capex build-out, rising compute and pricing dynamics, and why apparent productivity gains may be misleading. He also discusses physical bottlenecks, implications for monetary policy, and whether AI can meaningfully fix U.S. debt.
AI Snips
Chapters
Books
Transcript
Episode notes
Uncertainty Drove The AI Investment Cycle
- AI's macro importance comes from deep uncertainty that forces sustained investment cycles rather than immediate productivity wins.
- Steve Hou predicted a multi-year capex build as investors fund discovery because we couldn't falsify AI's potential early on.
Capex Buildout Is The Main GDP Channel
- AI's largest near-term GDP effect is from the capex buildout, not downstream productivity.
- Korea and Taiwan surged because chip and memory exports spiked to supply the US cloud/data center expansion.
Agentic AI Can Enlarge The Bubble
- The AI wave is manifestly a bubble but its duration and scale matter more than the label.
- Agentic AI (models calling models) can amplify compute demand by orders of magnitude, extending and enlarging the bubble.
