Amazon Discontinues Internal AI Usage Leaderboard Amazon has quietly ended an internal competition over who could consume the most AI, closing a popular leaderboard that had started to turn experimentation into a costly game rather than a productivity tool.

How the AI leaderboard emerged

In early 2026, Amazon employees built an informal internal dashboard called “KiroRank” to track and rank staff by the volume of AI tokens they used, reflecting a wider Silicon Valley trend of measuring AI enthusiasm by raw consumption rather than results. Tokens are the units large language models use to process text, and Amazon, like other big tech companies, closely tracks them because they directly drive cloud-computing costs.

As generative and “agentic” AI spread across the company, some employees reportedly began assigning autonomous AI agents to perform low‑value or unnecessary tasks simply to climb the rankings, a behavior dubbed “tokenmaxxing” (or “toxenmaxxing” in some reports).

Decision to shut KiroRank down

Concerns about rising compute bills and misaligned incentives prompted senior leadership to intervene. The Financial Times reported that Amazon “scraps AI leaderboard to stop workers chasing usage scores,” noting that costs were rising as workers tried to outdo one another on usage rather than solve real problems.

Dave Treadwell, a senior vice president, told staff: “Please don’t use AI just for the sake of using AI… Use AI to help you solve customer problems, to help you solve business problems, to innovate.” The Verge highlighted the same message, emphasizing Treadwell’s warning: “Don’t use AI just for the sake of using AI.”

Competing views inside Amazon

Amazon confirmed to Business Insider that KiroRank had been “deprecated,” stressing it was an employee-made experiment that “was never intended to promote the use of AI for usage’s sake.” A company spokesperson said Amazon wants AI adoption tied to “innovation and operational efficiency gains,” and that teams are free to choose how to use and track AI tools so long as they avoid pure tokenmaxxing.

The episode reflects a broader industry pivot: from celebrating raw AI usage to scrutinizing whether rapidly escalating AI budgets are delivering measurable productivity and business impact.

Continue reading https://foxvector.com/stories/019e7603-81fb-2036-715f-0497f3893907