In One Timeline

AI Keeps Getting Cheaper — So Why Is NVIDIA Exploding?

Why cheaper AI compute drives more demand, not less (Jevons paradox) — and NVIDIA's data-center surge.

1credit4sourcesAll episodes
Source pages are linked directly so license details can be verified at the original host.
Used For
Credit
Source
01Four NVIDIA H100 data-center GPUs
极客湾GeekerwanNVIDIA H100 (极客湾Geekerwan) 022.png / CC BY 3.0
commons.wikimedia.org

Data sources & references

01Stanford HAI — AI Index 2025 (cost per token at fixed capability)
02Epoch AI — training vs inference compute & scaling rates
03NVIDIA Investor Relations — Data Center segment revenue
04IEA — Electricity 2024 & Energy and AI 2025 (data-center demand, projections)