Cheaper AI tokens are driving more demand, and that's Jensen Huang's best-case scenario
A16z data shows a Jevons paradox in AI infrastructure: cheaper inference tokens are driving demand faster than unit costs decline, keeping GPU utilization and rental prices high. This dynamic benefits chip makers and cloud providers as long as token consumption growth outpaces price cuts—but a demand plateau would break the chain.
2026-10-11
Read full story



