The Economics of AI
Trillion-dollar capex against fast-falling token prices.
You're looking at a concept explainer: what this term means in plain English, and the records it ties together.
The industry's core tension: infrastructure spending measured in hundreds of billions per year, against inference prices that fall roughly tenfold per capability tier per year and open models pressing margins from below. Bulls point to genuine revenue curves (OpenAI and Anthropic reached multi-billion run-rates at record speed) and productivity gains; bears to depreciation schedules on GPUs and circular vendor-financing deals. Every stock on the Stocks page is a position in this argument.
How this connects to the rest of the stack
This record's links at a glance — click a group to jump to it.
Related
Sources
Epoch AI — data on AI economics ↗ · Sourced figures on this site's Trends page ↗
Record reviewed as of 2026-01-15 — see Methodology.