Meta internal "Claudeonomics" ranking revealed, 85k employees burn through 60 trillion tokens in 30 days

According to 1M AI News monitoring, a ranked AI usage leaderboard called “Claudeonomics,” named after Anthropic’s flagship product Claude, appeared on Meta’s internal network. It was built by employees using company data, compiled token consumption for more than 85k people, and listed the top 250. A copy of the leaderboard seen by The Information shows that, over the past 30 days, total consumption surpassed 600 trillion tokens. Using Claude Opus 4.6’s publicly disclosed average price (about $15 per million tokens) as a rough estimate, it comes to about $900 million. However, Meta’s actual model mix and protocol prices are unknown. The top-ranked individual user averaged 281 billion tokens consumed, and the cost could be as high as several million dollars.

The leaderboard includes gamified incentives, rising step by step from bronze to jade. The highest titles include “Token Legend” and “Session Immortal,” as well as “Model Connoisseur” and “Cache Wizard.” Some employees, to chase the rankings, have their AI agents run research tasks continuously for hours—purely to drive up usage. Neither Meta CEO Mark Zuckerberg nor CTO Andrew Bosworth made the top 250.

A recent trend in Silicon Valley is “tokenmaxxing,” where token consumption is becoming a new metric for measuring engineers’ productivity. In February this year at a technical conference, Bosworth said that a top engineer’s token spending had already been equal to their salary, with productivity improving as much as 10x. “It’s a sure win—keep burning, no limit.” Last month, Nvidia CEO Jensen Huang said that if an engineer earning an annual salary of $500k consumes fewer than $250k in tokens each year, he would be “deeply on alert.”

Meta engineers currently use external models from Anthropic, OpenAI, and Google, as well as internal tools like MyClaw (Meta’s version of OpenClaw) and the recently acquired Manus. In an internal memo this year, Zuckerberg made a “bold request” to the engineering team: rewrite Meta’s codebase so that AI agents can directly read and modify the code.