Amazon has reportedly discontinued an internal AI usage leaderboard after employees began competing to maximize AI consumption rather than focusing on meaningful productivity gains. The decision highlights a growing challenge facing companies worldwide: while AI promises efficiency and innovation, uncontrolled usage can also create significant costs without delivering proportional business value.

The internal system, known as “KiroRank,” was designed to encourage employees to adopt AI tools. Instead, it reportedly led to a phenomenon dubbed “tokenmaxxing,” where workers increased AI activity simply to improve their rankings.

What Was KiroRank?

KiroRank was an internal leaderboard that tracked employee usage of Amazon’s AI development tools. The objective was to accelerate AI adoption across teams and encourage experimentation with emerging AI technologies.

However, reports suggest the system unintentionally incentivized employees to focus on usage metrics rather than business outcomes. Some workers allegedly launched AI agents and automated tasks with limited practical value simply to increase their token consumption and climb the rankings.

As AI models process more data, companies incur higher computing costs. The excessive AI activity reportedly drove up infrastructure expenses without necessarily improving productivity.

Rising AI Costs Become A Reality Check

The episode reflects a broader concern that analysts have been raising for months. While companies continue investing billions into AI infrastructure, many are still struggling to measure clear returns on those investments.

Generative AI systems operate using “tokens,” which represent units of text and data processed by AI models. Every interaction consumes computing resources, making large-scale AI deployments expensive. As organizations expand AI usage across thousands of employees, operational costs can increase rapidly.

The Amazon case demonstrates how poorly designed incentives can turn AI adoption into a cost center rather than a productivity tool.

Amazon Says AI Should Solve Problems, Not Win Rankings

According to reports, Amazon leadership decided to remove the leaderboard after recognizing that employees were optimizing for AI usage rather than meaningful outcomes. Senior executives reportedly emphasized that AI should be used to solve customer and business problems, not simply to generate higher usage numbers.

Amazon has clarified that while it continues to monitor AI usage for operational purposes, it does not want employees using AI merely to improve metrics. The company is now reportedly shifting toward performance indicators tied to actual deployments and business impact rather than raw AI consumption.

Bigger Questions For The AI Industry

The incident arrives at a time when major technology companies are spending hundreds of billions of dollars on AI infrastructure, data centers, chips, and software development. Investors are increasingly asking whether these investments will generate sustainable returns.

Experts note that successful AI adoption requires clear business objectives, proper governance, and metrics focused on outcomes rather than activity. Without those safeguards, organizations risk creating expensive AI programs that generate impressive usage statistics but limited real-world value.

The Amazon episode may serve as an early lesson for companies rushing to integrate AI into every aspect of their operations.

Summary

Amazon has reportedly shut down its internal AI leaderboard, KiroRank, after employees began competing to maximize AI usage rather than focusing on productivity. The practice increased computing costs and highlighted the financial risks of poorly structured AI adoption programs. The incident underscores a growing industry concern that AI investments must be measured by business outcomes and efficiency gains, not simply by how much AI is being used.