When most people in crypto hear “token crackdown,” they reach for their hardware wallets. This time, the tokens in question have nothing to do with blockchain. Meta is imposing limits on how much its employees can use internal AI tools, according to a report from The Information, after discovering that its workforce had developed what can only be described as a competitive addiction to large language model usage.
The numbers are genuinely staggering. Meta employees collectively consumed 60 trillion tokens in a single 30-day period, with individual users burning through hundreds of billions of tokens on their own. To put that in perspective, a typical ChatGPT conversation might use a few thousand tokens. Some Meta employees were apparently treating AI tools less like productivity software and more like an all-you-can-eat buffet with no closing time.
The Claudeonomics era
An employee-built internal dashboard called “Claudeonomics” tracked token consumption across the company and awarded titles based on usage volume. The top tier? “Token Legend.” The dashboard created what amounted to an internal competition, with employees actively monitoring leaderboards to see who could consume the most tokens.
