Meta may introduce limits on AI costs for engineers

Meta's management is planning to introduce special limits on the use of AI resources for its employees in the future. According to an interview with Instagram head Adam Mosseri, within the next year or two, the amount engineers spend on AI tokens could equal their annual salaries. This implies that strict control mechanisms are necessary to maintain the company's financial stability. This is reported by Techcrunch.com reports .

Requests sent to AI systems and the responses received from them are calculated using a specific number of "tokens." Behind every token lies immense computing power and electricity costs. Speaking on Lenny’s Podcast, Mosseri noted that the "burn rate" of a high-performing engineer is becoming very expensive for the company. Therefore, it is expected that each employee will be allocated a certain token budget based on their work efficiency.

Tech giants are optimizing costs

Meta has already begun controlling AI costs internally. Specifically, the company shut down a "leaderboard" tracking employee token consumption. This was because it was discovered that the system could cost the company several billion dollars by 2026. According to Mosseri, there are currently no restrictions for Meta employees, but it is time to put an end to inefficient projects that act as "token burners."

Not only Meta, but other major IT companies are facing similar problems. For example, according to ixbt.com, Uber had already exhausted its AI coding budget for 2026 by April of this year. Microsoft, in order to save costs, canceled the Claude Code licenses provided to its engineers and moved everyone to its own Copilot CLI tool.

A new era of resource management

Adam Mosseri compared managing AI tokens to standard payroll or operating expenses (OpEx). He believes that just as resources like GPU, CPU, storage, and RAM are limited, AI capacity must also be distributed economically. In the future, the budget allocated to each engineer will be proportional to the return on investment (ROI) they bring to the company.

Nevertheless, Mosseri believes that in the long term, the cost of AI services will decrease. He predicts that as new AI models enter the market and competition intensifies, a "price war" will begin. This will open the door for companies to provide their engineers with even broader capabilities. For now, giants like Meta are forced to protect their budgets by cutting "mindless" spending.