Every time a user interacts with AI, it costs tokens, which are the basic unit of AI consumption.

OpenAI API tokens initially cost approximately a fifth of a penny per thousand tokens.

Token output at large companies like Meta and Microsoft is measured, sometimes via leaderboards or tools showing money spent on AI tokens.

A simple AI chat typically costs a few hundred tokens.

Computationally expensive AI models like Claude Mythos can only be used effectively by individuals who spend significant amounts on tokens.

The entire AI investment cycle is predicated on token demand.