Microsoft Cracks Down on 'Tokenmaxxing' Amid Strong Earnings

Microsoft has taken steps to curb 'tokenmaxxing', a trend where companies use AI tools extensively, often without delivering commensurate productivity gains. According to an internal memo by Jay Parikh, EVP of Microsoft's CoreAI org, employees have been instructed against maximising AI use and limits on how much engineers can spend on AI tools at work. The move comes despite strong earnings for the company.

The issue is not financial constraints, but rather wasteful AI usage. As Parikh said in an email to employees, 'Tokenmaxxing is not what we are optimizing for.' Microsoft has also made OpenAI's GPT-5.6 Sol the default AI model for internal use and introduced new limits on how much engineers can spend on AI tools.

This move is part of a growing trend among tech companies to curb AI token spend. Meta recently imposed its own token budgets for employees, and other companies such as Amazon, Adobe, Atlassian, and Citi have also looked to rein in employees' AI usage.