Why Microsoft is cracking down on ‘tokenmaxxing’ despite strong earnings

New DelhiUpdated: Aug 5, 2026 01:59 PM IST Satya Nadella during a discussion on the future of AGI, dismissed the idea of a single model ruling the world. (Express Image/Reuters)Microsoft has joined a growing list of tech companies looking to curb the wasteful use of AI tools internally amid greater scrutiny of the technology’s ballooning costs. The Windows-maker has instructed employees against maximising AI use and introduced new limits on how much its engineers can spend on AI tools at work, according to a report by 404 Media. Microsoft employees have also been asked to review their individual spend. The new directives came in the form of an internal memo issued earlier this week by Jay Parikh, the EVP who runs Microsoft’s sprawling CoreAI org.Microsoft has also made OpenAI’s GPT-5.6 Sol the default AI model for internal use. So far, Microsoft’s GitHub Copilot setup for internal use functioned as an auto-router that defaulted primarily to Anthropic’s AI models. As a result, Microsoft engineers were mostly generating code with expensive Claude tokens. A token is a unit of text processed by the AI model. Additionally, GitHub Copilot, which is an AI coding tool owned by Microsoft, was running at extremely negative gross margins before it moved to usage-based billing earlier this year, according to tech newsletter Sources. “As we accelerate our use of GitHub Copilot to deliver on our goals, we all need to be aware of how we consume tokens,” Parikh was quoted as saying in an email to employees. “Tokenmaxxing is not what we are optimizing for. I want all of us focused on maximizing outcomes that move the needle for our customers and our business. As such, we are updating our internal guidance and managing token spend with the same discipline we apply to every other critical resource,” he added. Parikh further said that “shifting more workloads to OpenAI models helps us get greater value from our token investment.”Story continues below this ad Last month, Microsoft updated its internal GitHub Copilot guidelines to state that every division of the company will have an AI token budget target. “While there is no target spend value being shared at this time. The data shows that many engineers spend in the range of hundreds of dollars a month to a few thousand dollars in tokens,” the guidelines reportedly said. It also mentioned that some decisions may place further restrictions as Microsoft monitors AI spend. What is tokenmaxxing? Why is Microsoft cracking down on it? Since the start of the year, several big tech companies have urged employees to use AI tools as extensively as possible. This trend was referred to as ‘tokenmaxxing’. Meta, for instance, even evaluated employees’ performance based on their usage of AI tools. However, the price for using AI tools has increased substantially in the last few months while not always delivering commensurate productivity gains. Several medium and large organisations that incorporated AI tools into their workflows to boost productivity began to question whether using AI tools translates to creating real value. The disconnect between rising AI adoption and stagnant revenue growth may stem from the fact that tools like Claude Code have dramatically lowered the cost of producing code, but not necessarily the cost of producing meaningful outcomes.Story continues below this ad To be sure, Microsoft’s recent move to end tokenmaxxing is likely not influenced due to financial constraints. In fact, its latest earnings report shows the company revenue, operating income, and net income all increased for the year, beating Wall Street expectations. Based on Parikh’s email, the change in Microsoft’s policy on AI spend is aimed at curbing wasteful AI usage. “We are not optimizing for fewer tokens. We are optimizing for more impact per token,” he said, while emphasising that Microsoft will continue moving toward becoming ‘AI-first’. What have other tech companies done? Microsoft is far from the firs