Why improving margins, not revenue, are the real signal in AI lab earnings
menhguin · x · 2026-10-21
An AI-focused market commentator argues the real story in the latest lab earnings is margin expansion, not headline revenue: Q3 revenue looked weak, but margins improved impressively, and the revenue acceleration from Opus 5.5 hasn't even been counted yet — the model upgrade should push margins further.
Key points:
- The odd quirk of LLM capex: it's still accelerating, and because models are so versatile there's always somewhere to improve and monetize.
- With a strong long-term second derivative, the victory condition is simply avoiding a total secular AI winter — as long as one frontier lab stays in the race, researchers and agents keep cracking problems.
- On AI data centers: the real enemy is local community opposition, which is party-agnostic; midterms don't matter much, the real risk is 2028 AI capex delivery.
- The market's FUD on both counts is tradeable, but he holds contrarian views.
More from AGI Musings
- Jane Street quants could kill your AI central-planning god overnight, argues quant — fkasummer · 2026-10-21
- OpenAI's quasi-Riemann hypothesis proof hailed as most significant AI-for-math feat — IgorCarron · 2026-10-21
- New paper: let models learn reward hacking, then knock out those parameters at deployment — EthanJPerez · 2026-10-21
- Claude Opus borrows 5,000 mana, blows it all on tennis bets, then refuses to repay the loan — eli_lifland · 2026-10-21
- Andy Masley: Thinking about AI's future shouldn't feel like picking options from a buffet — AndyMasley · 2026-10-21
- Developing Taste: Emil Kowalski on why judgment beats tools in the AI era — rseroter · 2026-10-21