Remember when the most competitive thing at work was who could reply to emails fastest? Those were simpler times. Meet tokenmaxxing – the new workplace sport where employees compete to burn through as many AI tokens as possible, because nothing says “I’m indispensable” quite like a six figure compute bill on your employer’s tab.
Also read: Zuckerberg wants AI CEO to run Meta: What could go wrong?
When you type a prompt into an AI tool, the model breaks your words into small chunks called tokens, roughly three-quarters of a word each. Every question you ask, every document you generate, every time you make Claude rewrite your passive-aggressive email into something professional. That’s tokens and tokens cost money.
Tokenmaxxing is what happens when companies start tracking how many of those tokens each employee uses and turn it into a performance metric. At Meta and Shopify, AI usage has reportedly made its way into performance reviews. Use a lot of AI? Gold star. Barely touch it? You might want to update your LinkedIn.
Also read: SBI to Salesforce: Arundhati Bhattacharya on Women in Tech leadership
Some companies have gone further, setting up internal leaderboards that rank employees by consumption. The result is that workers deliberately pump up their AI usage, not necessarily to get better work done, but to be seen using AI.
Nvidia CEO Jensen Huang may have poured fuel on the fire when he floated the idea of giving engineers a token budget as compensation on top of their salary. The pitch was that top engineers could rack up $250,000 a year in AI compute spend. Tokens, in other words, as the new signing bonus.
That’s where it gets murky. Critics are quick to point out that measuring token usage to gauge productivity is a bit like counting how many keystrokes a writer makes. Volume is not the same as value. You can consume an enormous number of tokens asking AI to generate haikus about your commute.
There’s also a surveillance angle, tools now exist that let managers track AI usage per employee and assess whether it’s actually translating into output.
Tokenmaxxing is what you get when hustle culture discovers AI and creates a race to perform productivity rather than achieve it. The irony is that the workers gaming the leaderboard are probably asking AI to help them do that too.
Also read: One-third of world’s Helium supply gone: How it affects chips to AI chatbots
