Will scale economies produce a few dominant firms, or will open models and falling inference costs commoditize intelligence?
Scale economies, data feedback loops, and capital walls point to oligopoly; open weights and collapsing inference prices point to commodity. Competition authorities are watching the vertical stack, not just the model market.
View on the map → · Open in Browse →
The question split by segment: Lambert's evidence that the open-closed gap is stable rather than closing, and his forecast of closed oligopolies in premium agents with open models winning a low-margin commodity market. Two new structural forces entered: a likely compute crunch, and de facto US model licensing as a regulatory entry barrier.
Open models are not meaningfully accelerating towards matching the best closed models in absolute performance.
Evidence the ~six-month open-closed gap is stable; open models matter through specialization and diffusion, not frontier parity.
Closed labs become premium integrated-agent oligopolies; open models win a larger, low-margin commodity market.
A compute crunch is likely near—particularly for the long-context workloads that drive agentic AI.
Token demand growing ~10× yearly against 3.4× supply growth challenges the cheap-inference commoditization assumption.
Although training the world's best AI model is enormously expensive, training one that is just as good—but six months later—is cheap.
Fast-following and RL's declining reliance on user data undermine natural-monopoly claims.
Government approval is now effectively required for frontier releases — a regulatory entry barrier independent of scale economies.