Microsoft Halts Tokenmaxxing with Strict Budget Caps

Microsoft has imposed strict budget caps on its employees' use of AI tokens, following a trend of 'tokenmaxxing' that has spread across Silicon Valley.

The company has designated GPT-5.6 as its default internal model due to its lower costs, and employees can view their personal AI token consumption in real-time.

Much of the issue stems from the practice of using AI for the sake of using it, rather than focusing on actual productivity gains.

'Tokenmaxxing' refers to the distorted incentive mechanism that treats AI usage as an internal KPI for enterprises.