A Theory of Training Profit-Optimal LLMs

In three linesEconomic model combining scaling laws and microeconomic theory to characterize rational behavior of LLM training firms. Analyzes profit maximization under compute-bound and data-bound regimes: in compute-bound, optimal model size tracks hardware efficiency (FLOPs/$) at near-linear rate; in data-bound, optimal training expenditure scales as D²/E.

Summary generated by Claude — human-verified