Amazon Tells Engineers to Stop 'Tokenmaxxing' AI Usage Stats

Amazon published updated engineering guidelines on October 27 that order its developers to stop inflating AI token usage to appear productive, a practice the company calls "tokenmaxxing." The message is blunt: consuming more tokens is not the same as delivering more value. The caveat matters as much as the rebuke. Amazon is not retreating from AI. "We are certain that whatever future best practices are, AI will be at the center of software engineering practice," the company wrote in the memo.

What does the new guidance change?

The guidelines target employees who deliberately pump up AI usage on internal dashboards. According to the memo, metrics such as AI adoption rates, token consumption, code changes, and code reviews become "deeply flawed" when they are disconnected from actual results. Turning those metrics into targets, Amazon argues, encourages wasteful behavior. That waste carries a real cost. The company says the practice clashes with its stated emphasis on "frugality and simplicity" and can raise both compute bills and software complexity.

Amazon's preferred yardstick is patience, not dashboards. "We believe that it is best to measure, and be measured by, long-term outcomes," the company wrote. Managers should not discard usage data entirely. Input metrics remain useful when examined "with care and judgment" alongside other indicators. Amazon suggests managers examine both the teams using AI the least and those using it the most, to determine whether adoption actually translates into better delivery.

The memo also pushes back on a quieter assumption: that every task needs an AI tool. If current tooling does not benefit someone's work, Amazon says, that is "a finding, not a failing." Engineers who can explain where today's tools fall short are demonstrating judgment, not reluctance.

Why now?

In May, it emerged that Amazon employees were using AI unnecessarily to boost internal usage scores, amid pressure to adopt the company's own tools. Workers reportedly used MeshClaw, an internal AI agent that can deploy code and interact with Slack, to inflate their token counts. According to CNBC, Amazon shut down an internal leaderboard that same month after workers were caught gaming it to climb the rankings — an early instance of exactly the behavior the new guidelines now discourage. Notably, the company says it has not imposed token caps on employees.

The push comes as Amazon, under CEO Andy Jassy, embeds AI deeply across its operations. Jassy said earlier this year that the technology will "change the scope and speed" of everyday work. He has also suggested automation could shrink the company's corporate workforce over the next few years, though he expects it to create "other jobs." That workforce claim remains a suggestion, not a confirmed plan.

Is this an Amazon problem alone?

No. Uber president and COO Andrew Macdonald said in May that higher token usage has yet to translate into useful consumer features. "That link is not there yet," he said. The remark carries weight because Uber's CTO previously revealed the company had burned through its entire 2026 Claude Code budget by April.

Accenture appears to be wrestling with the same problem on behalf of clients. Leaked audio obtained by 404 Media in June captures executives acknowledging that token spending is climbing fast as companies move from simple chatbots to agentic workflows, and that nobody can reliably predict what a given task will cost or whether it was worth the money. Accenture's agentic AI strategy lead described it as "not a niche problem." The firm is reportedly building an advisory tool called "Token IQ" to help customers manage the bill. OpenAI CEO Sam Altman has previously acknowledged that AI token costs had become "a huge issue" as companies faced mounting bills from heavy use.

Amazon's guidelines leave one question open. The company asserts AI will sit at the center of software engineering, yet its own memo admits current metrics cannot prove that claim — and the industry-wide bill, from Uber's exhausted budget to Accenture's unbudgetable agent costs, keeps climbing while the measured value lags behind.