By Isabelle Bousquette

Is maximizing AI usage inside a company always a good thing?

That's the question startups, investors and big corporations were asking after an internal dashboard at Meta Platforms went viral for ranking employees by their individual token usage and doling out flashy titles like "Token Legend."

Tokens refer to the quantifiable chunks of information that AI models process. OpenAI estimates one token is about four characters.

The Meta dashboard, a side project created by an individual employee first reported on by the Information, has since been taken down at the employee's discretion, a Meta spokesperson said. Meta has a separate AI Insights dashboard that tracks AI usage more holistically than just by tokens used.

But the publicity around the leaderboard has stirred up a vitriolic debate about the merits of "tokenmaxxing," a Silicon Valley term for using as many tokens as possible.

"Outcome maxxing >> token maxxing," Yamini Rangan, chief executive of software company HubSpot wrote on LinkedIn, expressing a common sentiment that maximizing AI usage is pointless unless it has real returns. Rangan and others argued that pure token use is the wrong metric to optimize for.

"You could tokenmaxx all day long but get outcomes that are not what you desired," said Andrew Lau, co-founder and CEO of AI software engineering company Jellyfish.

Brian Elliott, co-founder and CEO of enterprise AI code generation firm Blitzy said tokenmaxxing is like trying to measure a company's revenue based on the number of cold calls its sales team makes. "You need to take a more deliberate approach."

But some companies disagree, arguing that the tokenmaxxing mindset is actually critical to their survival.

Corporate America remains steadfast in the belief that companies who don't fully learn to leverage AI will get outcompeted and become obsolete. And while leaderboards and incentive systems aren't perfect, executives say they're effective and driving more internal AI adoption. Which they say is worth it -- even if it costs the company hundreds of thousands of dollars, and even if some of the AI usage ends up being worthless.

"It is existential for us," said May Habib, CEO and co-founder of enterprise AI startup Writer, about driving more internal AI usage. "We are in the most competitive space that has ever existed and will ever exist."

Writer has its own internal token consumption leaderboard -- which Habib makes clear to employees that she monitors -- in order to encourage employees across the company to use more AI. There are no penalties for being at the bottom, but those at the top get a round of applause in Slack, and sometimes a gift card.

Leaders from the month of March include an employee who used nearly 11 billion tokens that month, followed by one who used just over 6 billion.

The cost of tokens can vary depending on which provider is used. Employees at Writer use their own internal AI platform, where 10 billion tokens is just north of $50,000.

Habib knows token consumption is a gameable metric. She knows some employees will use tokens for personal projects. She knows that not all the tokens will end up driving value for the business. But she's OK with all of that.

"The second you start thinking about the individual business [return on investment] of one agentic action, you will never do anything agentically," she said. The broader goal is a mindset shift. And offering incentives to employees with a leaderboard is effective, she said.

At AI customer support company Sendbird, a token consumption leaderboard has been "super-impactful," according to growth marketing lead Abhi Jothilingam. "People are motivated by competition," he said.

Since the leaderboard has been rolled out, Jothilingam said he tests out more ideas for AI apps and builds them faster. He's vibe coded a social media tracker and a project management tool, shooting him up to a rank of No. 5 on the leaderboard at the firm, which has a couple hundred employees.

The Wall Street Journal reported Sunday how some AI companies were scuttling products and even applying limits to the amount of tokens available, citing a computing capacity crunch. Yet many companies remain undeterred.

"We all should be tokenmaxxing," said Sonya Huang, a partner at Sequoia Capital. Huang said Sequoia has its own leaderboard, as do many of the portfolio companies she advises. Sequoia also offers firmwide AI office hours with the engineering, product and design teams to help drive more internal AI usage.

Critics of tokenmaxxing are missing the point she said. Yes, it's an imperfect metric, she admitted, but "the thing that matters for your company is: is my employee becoming insanely AI-pilled? And that requires getting them on this tokenmaxxing mindset."

"There is this insane new piece of technology that is fundamentally going to rewire how we work. Some companies are going to make it, some companies are not," Huang said. "So how can you psychologically gamify the experience of just getting everybody on board this evolution as quickly as possible?"

Write to Isabelle Bousquette at isabelle.bousquette@wsj.com

(END) Dow Jones Newswires

April 14, 2026 08:00 ET (12:00 GMT)

Copyright (c) 2026 Dow Jones & Company, Inc.