Tokenmaxxing
The rise and fall of Silicon Valley's latest productivity metric.
First we got looksmaxxing, then sleepmaxxing, and then came tokenmaxxing, or maximising the number of tokens used by AI agents as a proxy for productivity. At Meta, an employee ran an internal leaderboard called “Claudeonomics” that ranked employees by token usage and handed out titles such as “Token Legend” and “Session Immortal” to the heaviest users.[1] At Salesforce, engineers faced minimum token-spend targets and a desktop widget that refreshed their usage every fifteen minutes.[2] Nvidia CEO Jensen Huang said in March 2026 that he would be “deeply alarmed” if a high-level engineer earning $500,000 a year was not consuming at least $250,000 worth of AI tokens a month. He compared it to a chip designer refusing to use CAD tools and choosing to work with only paper and pencil instead.[3]
At first, increasing token usage was framed as a way to encourage developers to try new AI tools and surface new use cases. Shopify’s Head of Engineering, Farhan Thawar, said on The Pragmatic Engineer Podcast in June 2025: “We have a leaderboard where we actively celebrate the people who use the most tokens because we want to make sure they are [celebrated] if they’re doing great work with AI.”[2] Elsewhere, dashboards like this were quickly, and not unexpectedly, gamified, with employees in some companies setting agents to do pointless work simply to rise up the rankings.[2] As the cost of token usage ballooned without clear productivity gains, some companies began to pull back. Amazon shut down its internal ranking system; senior vice-president Dave Treadwell told staff not to “use AI just for the sake of using AI.”[4] At Uber, which exhausted its full 2026 AI budget in the first four months of the year, COO Andrew Macdonald said he had seen no evidence that higher token consumption produced more productivity: “That link is not there yet.”[5]
Sources:
[1] Mann, J. (6 April 2026). Meta Employees Vie for AI ‘Token Legend’ Status. The Information. https://www.theinformation.com/articles/meta-employees-vie-ai-token-legend-status
[2] Orosz, G. (23 April 2026). The Pulse: ‘Tokenmaxxing’ as a weird new trend. The Pragmatic Engineer. https://blog.pragmaticengineer.com/the-pulse-tokenmaxxing-as-a-weird-new-trend/
[3] Huang, J. All-In Podcast, recorded live at Nvidia GTC 2026 conference, San Jose (March 2026). See also: Business Insider (19 March 2026). https://www.businessinsider.com/jensen-huang-500k-engineers-250k-ai-tokens-nvidia-compute-2026-3
[4] Rosner-Uddin, R. (28 May 2026). Amazon scraps AI leaderboard to stop workers chasing usage scores. Financial Times. https://www.ft.com/content/b1a62a7f-6df5-4c90-94ce-64ce9c9961b6
[5] Angelo, J. (26 May 2026). Uber burned through its entire 2026 AI budget in four months. Now its COO is questioning whether it’s worth it. Fortune. https://fortune.com/2026/05/26/uber-coo-ai-spending-tokens-claude-code/