Meta's Adam Mosseri predicts AI token budgets may soon be capped per engineer

Instagram head Adam Mosseri predicted in a recent interview on Lenny's Podcast that technology companies may need to impose caps on AI token spending per engineer within one to two years, as the cost of processing AI prompts threatens to rival employee salaries as a major operating expense.

"I think that you can imagine, at least in a year or two … that the burn rate of a strong engineer might be the same as their salary, or their cost of employment. And in that world, you're going to probably need to put in some caps," Mosseri said. The remarks mark one of the first times a top executive at a major tech company has publicly acknowledged that AI inference costs could become a line item as significant as payroll.

AI token spend has become a hot topic in Silicon Valley after Meta was reportedly on track for billions of dollars in AI costs in 2026, prompting the company to shut down an internal AI token spend leaderboard. Mosseri noted that Meta has already reined in costs somewhat by shutting down "silly things" it was doing — like that token spend leaderboard.

Mosseri drew a direct analogy to how companies manage other finite resources. "I think of it like…any other resource," he said. "I have to decide how to deploy capacity to my different teams because I have a limited number of GPUs and CPUs and storage and RAM etc. I have to decide how to deploy OpEx for labeling budgets across my teams. I have to decide how to deploy payroll for headcount across my teams." Token budgets, he argued, will work the same way.

The cap per engineer would have to be proportional to the company's trust in their ability to use the budget in an "ROI-positive" way, Mosseri added. Meta does not currently have token caps for any employee, but Mosseri believes their use "could be healthy" in the future. Further down the road, he expects token costs to fall as AI model makers enter a pricing war to attract users.

Meta is not alone in grappling with AI cost overruns. Uber blew through its entire 2026 AI coding budget by April, triggering what the company called an "AI reckoning." Soaring token costs also led Microsoft to cancel Claude Code licenses, consolidating its engineers around its own Copilot CLI tool instead. These parallel episodes suggest that the entire industry is navigating uncharted cost territory.

"It's not that hard to build a token incinerator, and that doesn't create a lot of value," Mosseri said. His central argument — that AI tokens are a finite resource that must be managed like any other operational cost — signals that AI infrastructure governance is shifting from an era of unlimited experimentation to one of disciplined budget management, with implications for every company deploying AI at scale.