Why compute supply favors closed models first, and when open deployment wins
yangyi · x · 2026-09-28
A long-form take on compute supply economics, responding to discussion about open-source model deployment opportunities:
- During supply scarcity, identical machines and electricity go to closed models first — they pay more and acquire customers more easily, so commercial efficiency is higher
- Open-source deployment only scales once localized models mature and supply reserves (power, data centers) expand, since open models inherently carry lower margins
- Another angle: only after closed models fully monopolize pricing and supply-side premium disappears will providers pivot to open-model hosting
- The author also argues "models get cheaper" depends on your reference frame: what's valuable isn't the model itself but data-driven refinement — accumulated human and AI feedback about the world. Models capable of solving future-era problems will command very high prices
Related event: Debate over private deployment of open-source models(2 posts)→
More from AGI Musings
- Gwern's 2022 AI takeover short story hailed as eerily prescient on sandbox failures — PandaAshwinee · 2026-09-29
- AI evangelist returns home for Mid-Autumn Festival, family dismisses his AGI warnings — huangyun_122 · 2026-09-29
- The doctors who thrive won't resist or worship AI — they'll know exactly when it's wrong — realmeetjames · 2026-09-29
- 70% of Stripe API commands now come from AI agents as agent commerce surges — KyeGomezB · 2026-09-29
- The low-ceiling hypothesis: why AI scaling might end far below superhuman cognition — coherence · 2026-09-29
- Security veteran: AI is radicalizing exploit development like the worm era of 2000 — joshua_saxe · 2026-09-29