Forget about the number of lines of code written. Now IT professionals have a new way to compete for the status of the most productive employee: by the number of AI tokens used. Because of this, a new trend has emerged in companies called tokenmaxxing, which has already provoked heated debate.
Forget about the number of lines of code written. Now IT professionals have a new way to compete for the status of the most productive employee: by the number of AI tokens used. Because of this, a new trend has emerged in companies called tokenmaxxing, which has already provoked heated debate.
Armed with AI tools for writing code, developers have a whole "wallet" of tokens at their disposal. They began to be offered to engineers as a reward, and they also appeared in job descriptions at OpenAI and Anthropic, Business Insider writes .
What’s more, Meta engineers compete to spend tokens to get into a ranking called Claudeonomics, a dashboard that tracks AI usage and allows employees to compete for titles like “Token Legend.”
But is token spending a good indicator of developer productivity?
Some see this as a useful marker that employees are actively using new tools; others warn that it could encourage inefficient use of AI within companies, leading to a “show game” of metrics.
Tokenmaxxing is the pursuit of spending as many tokens as possible. Some tech companies have dedicated rankings of token leaders, such as Meta and OpenAI.
While it’s hard to gauge how widespread tokenmaxxing has become, corporate spending on AI is clearly on the rise. Fintech firm Ramp called it a “$1 trillion blind spot,” citing Gartner data that shows monthly business spending on AI has quadrupled in the past year.
And they’re also showing off. Startup founders and progressive engineers are posting their token spending on X to show how deeply immersed they are in AI. One xAI employee wrote that the tech community is turning every good idea into “theater.”
“I personally spend thousands of dollars on tokens every week… it seems crazy, but I can’t stop tokenmaxxing,” wrote a user under the nickname ben guo on X.
“Y Combinator CEO Harry Tan seems to be a fan of this idea. Under a post criticizing companies that “hoard” tokens, he noted, “We’ve been doing tokenmaxing longer than most people.”
Some in the tech world argue that tokenmaxxing is an effective metric; others call it a pointless waste.
Khosla Ventures partner John Chu called evaluating work by the number of tokens spent an “absolutely pointless policy” on social network X.
“Many of my friends at Meta have told me that because of this policy, people are creating bots that just run in circles, burning tokens as fast as they can,” he wrote.
Edwin V. Arbus, of Cursor, offered a more nuanced view of the metric, calling it “a useful and quick, if somewhat imperfect, guide.” He compared it to body mass index (BMI), which gives a general idea of health but doesn’t take into account muscle or bone mass.
While Nvidia CEO Jensen Huang didn’t directly comment on the term tokenmaxxing, he stressed the importance of engineers consuming large amounts of tokens. He said that if an engineer with a salary of $500,000 didn’t spend even $250,000 worth of tokens, it would cause him “deep concern.”
Gergely Oros, a contributor to The Pragmatic Engineer, called the practice wasteful. “Developers manipulate any metrics that are used as criteria for bonuses or promotions,” he wrote. “This case is no exception.”
Ben Puladian, founder of BEP Research, drew a different conclusion from this trend, calling computing power the main barrier to innovation. “In the era of AI, every employee becomes a consumer of computing resources,” he wrote on social media X.
“Token spend is always an output, not an input,” wrote Persona engineer Arush Shankar, who, according to his LinkedIn profile, previously worked at Square and Microsoft. “This is a metric to consider, but should never be viewed in isolation. It’s a signal, but it’s not a deciding factor.”
