Do Son 
According to a report from The Information, Meta is taking steps to limit how much artificial intelligence its employees use internally. Current usage levels are so high that the company expects costs to climb into the billions of dollars. For Meta, this represents an enormous expense. As a result, the company now plans to rein in AI usage before costs spiral out of control.
From Encouragement to Performance Metrics
Previously, Meta actively encouraged employees to adopt AI tools to boost productivity. To reinforce this push, the company even incorporated AI usage into some performance reviews. However, wherever performance metrics exist, workarounds inevitably follow. Indeed, posts on Reddit have discussed methods for quickly burning through large numbers of tokens to hit these targets.
A Shift Toward Token Minimization
Meta is now pivoting toward what it calls a token minimization strategy. Under this approach, employees should avoid using AI for tasks where it isn’t truly necessary, thereby reducing overall consumption. To enforce this shift, Meta plans to set AI budget caps for each team, establish token usage limits, and encourage employees to rely on the company’s own AI tools, such as Meta Code, to help control runaway costs.
According to the report, Meta employees consumed as much as 60 trillion tokens in just the past 30 days alone. AI spending has become a massive line item for the company. Looking ahead, Meta expects its total AI usage costs for 2026 to reach into the billions of dollars. Even a company of Meta’s scale considers these costs excessive, and management has concluded that usage must be brought under control.
Building Tools to Track AI Usage
Meta plans to begin real-time tracking of employee AI usage during 2026, alongside initial caps such as maximum quotas per team. By 2027, the company intends to implement dedicated structural management, including strict budget allocations and supporting tools. This effort will also involve building internal platforms capable of tracking employee AI usage and its associated costs in real time.
Not Just a Meta Problem
Meta is far from alone in tackling rising AI costs. Companies including Uber, Microsoft, and Salesforce have recently implemented similar restrictions. Across the industry, businesses are shifting away from unlimited, encouragement-driven AI adoption toward more disciplined cost management. Without such controls, AI spending risks becoming a significant operational burden. Whether these limits will improve efficiency remains uncertain, but one thing is clear: the massive expenses are real, leaving companies little choice but to manage AI usage carefully.
Support Our Threat Intelligence
If you find our CVE report and cybersecurity news helpful, consider supporting our work.
Buy Me a Coffee
PayPal
