-
Amazon shuts down employee-created leaderboards that promoted excessive use of AI.
-
An Amazon spokesperson said: “It is in no way intended to promote the use of AI for any purpose.”
-
Amazon’s move comes amid a broader review of “token maxing” and AI spending in general.
Amazon is trying to minimize “tokenmaxxing”.
The company will close employee-created leaderboards that track AI token usage. This was because some staff were encouraged to perform tasks that did not necessarily solve problems, just to move up the ranks.
“Don’t use AI just for the sake of using AI,” Amazon senior vice president Dave Treadwell told employees earlier this week. “We use AI to help solve customer problems, solve business problems, and innovate.”
The move, first reported by the Financial Times, is the latest sign that big companies are pivoting away from a culture of freewheeling AI spending amid concerns that AI budgets are ballooning and not getting a fair return.
An Amazon spokesperson confirmed to Business Insider that the company’s internal leaderboard, called KiroRank, has been “deprecated.” A spokesperson said the dashboard was an unofficial tracker created by a group of employees and was “in no way intended to promote the use of AI for exploitation purposes.”
“We are focused on implementing AI and sharing best practices, and celebrating innovation and operational efficiency across the company. We are proud that our teams are embracing this technology,” a spokesperson said in a statement.
A spokesperson said Amazon teams have the freedom to decide how to use the AI tools and how to track their usage. While the tech giants are tracking the usage of AI tokens to measure costs, they are not encouraging “token maxing,” a Silicon Valley trend that measures AI productivity by the amount of AI used by employees and companies.
That trend appears to be losing support.
For example, Uber Chief Operating Officer Andrew MacDonald recently said the ride-hailing giant hasn’t realized productivity or other gains commensurate with increased AI spending. McDonald’s comments came after Uber Chief Technology Officer Praveen Nepali Naga made headlines when she said the company had already exhausted this year’s Claude Code budget by April.
Tokens are how large-scale language models break down words into numerical inputs and outputs, and are essentially the building blocks of AI chatbots and coding tools. Token usage increased dramatically in 2026. One reason for this is the rise of agent AI, which allows agents to operate continuously, sometimes for hours on end, with minimal human intervention.
