For the past two years, the message from Big Tech to its employees was simple: use more AI. The companies that used it most would win. The numbers would follow.
The numbers have not followed. And now the message is changing.
What Amazon did and what it told employees
Amazon (AMZN) shut down an internal AI leaderboard called KiroRank on May 29, which had been tracking AI token usage among employees on the company’s internal Kiro developer platform, according to Business Insider, which confirmed the shutdown with an Amazon spokesperson. The story was first reported by the Financial Times.
More AI:
- Micron sits at the center of a red-hot chip rally
- IBM CEO sends blunt message on AI and quantum computing
- Anthropic CEO makes shocking admission about AI
Dave Treadwell, Amazon’s senior vice president of engineering, addressed the issue directly. “Please don’t use AI just for the sake of using AI,” he told staff. “Use AI to help you solve customer problems, to help you solve business problems, to innovate.”
An Amazon spokesperson confirmed the leaderboard had been “deprecated” and said the dashboard was an informal tracker created by a group of employees and “was never intended to promote the use of AI for usage’s sake,” Business Insider confirmed.
What KiroRank was and why it backfired
KiroRank scored Amazon employees on AI activity on the Kiro developer platform. Amazon had set a target for more than 80% of its developers to use AI tools weekly, according to Fortune.
Tokens are the units large language models use to process text and generate responses. Higher usage means higher compute costs.
What happened next was predictable in hindsight. Employees began inflating their scores through tokenmaxxing: running meaningless tasks through AI agents to consume tokens and climb the rankings.
Costs rose. Business value did not.
Amazon has now shifted from raw token counts to a metric it calls “normalised deployments,” which measures AI-assisted code that actually ships rather than token consumption alone, according to Let’s Data Science.
Nagle/Getty Images
Why Amazon is not alone in this reversal
Amazon’s retreat from tokenmaxxing metrics is the latest in a series of similar moves across the industry. The same week, Meta also abolished its own internal AI usage leaderboard, called Claudenomics, which had tracked token usage across 85,000 employees and singled out the top 250, according to Fortune.
Uber’s COO Andrew Macdonald recently said the company has not found a clear relationship between increased AI spending and the delivery of successful products. His comments followed a disclosure that Uber had already exhausted its annual budget for certain AI coding tools by April.
Microsoft canceled Claude Code licenses across its Experiences and Devices division earlier this month, citing cost concerns.
The pattern across Amazon, Meta, Uber, and Microsoft is consistent: heavy AI usage generated costs but not proportional outcomes, and leadership has begun to say so publicly.
Key figures on Amazon’s KiroRank shutdown and the broader tokenmaxxing backlash:
- Amazon: KiroRank leaderboard shut down May 29; tracked AI token usage on internal Kiro developer platform; replaced by “normalized deployments” metric measuring AI-assisted code that actually ships, according to Business Insider
- Treadwell’s message: “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,” Business Insider confirmed
- Context: Amazon targeted 80%+ weekly AI usage among developers; it is spending approximately $200 billion in 2026 capital expenditure primarily on AI and data centers, according to Fortune
- Meta: shut down Claudenomics leaderboard the same week; had tracked AI token usage across 85,000 employees and highlighted the top 250.
- Uber: COO said company found no clear link between AI spending and product delivery; had exhausted its annual Claude Code budget by April
- Microsoft: canceled Claude Code licenses across its Experiences and Devices division citing cost concerns; redirected engineers to GitHub Copilot CLI
What this signals for AI spending discipline across Big Tech
The leaderboard shutdowns are a symptom of a deeper problem that has been building since generative AI tools became standard in enterprise environments.
Token-based pricing means costs scale with usage, not outcomes. When companies incentivized usage without measuring results, they created a system where inflating numbers was rational for individuals but destructive for budgets.
The shift toward outcome-based metrics is a necessary correction. Amazon’s “normalized deployments” framework, code that ships rather than tokens consumed, is the kind of accountability layer that should have been in place from the start. The fact that it is arriving now, after billions in AI infrastructure spending, is the more revealing signal.
For investors watching Big Tech AI spending commitments, the message is nuanced. Amazon is not stepping back from AI. Its $200 billion 2026 capital expenditure commitment is intact.
But the internal discipline it is now applying suggests the next phase of the enterprise AI story will be defined less by adoption headlines and more by who can show returns on the infrastructure they have already built.
