Azeem Azhar: no sign of oversupply to paying customers despite AI bubble fears
2C_ornot2C · x · 2026-10-21
Responding to Carlota Perez, Azeem Azhar argues why the current AI cycle differs from past manias (GFC, railroads):
- The bear case is the easiest to make — apply historical mania patterns, cite CAPE/Shiller vs P/E divergence — but past divergences haven't always led to busts (e.g., the Australian mining boom), a fact often ignored.
- The fundamental difference: no sign of oversupply of capacity for customers willing to pay. His team controlled for circular customer behaviour (venture-backed companies buying from each other) and sees a more distributed customer base with real corporate spending.
- Details in his Substack AI Investment Brief.
Related event: Azeem Azhar on the AI Bubble: Real Productivity Boom, Wobbly Financing(2 posts)→
More from AGI Musings
- Memia #2026.42: AI agents scan 3,830 articles to curate 312 future signals — ben_r · 2026-10-21
- No frontier model is honest enough to enable recursive self-improvement, researcher says — scottleibrand · 2026-10-21
- X thread argues the 'death of hobbies' complaints are really about lost superiority — kevinnbass · 2026-10-21
- AI unicorn founders' average age drops from 40 to 29 as startups hit unicorn in 4.7 years — FinanceYF5 · 2026-10-21
- Schmidt and Ferguson Argue China and the U.S. Are Building "Empires of AI" — marigo · 2026-10-21
- Reviewer laments ICLR review motivation is gone as AI erases respect for authors' effort — AvivTamar1 · 2026-10-21