Meta Executive Says AI Token Spending Limits Could Become Necessary for Engineers

Meta may eventually place limits on how much artificial intelligence computing individual engineers can use as the cost of running advanced AI systems continues to climb, according to Instagram head Adam Mosseri. Speaking on Lenny’s Podcast, Mosseri said AI token spending could become significant enough within the next year or two for companies to manage it like any other business resource.

AI token spending refers to the cost of processing prompts and responses through AI models. As businesses expand their use of generative AI tools for coding, research, and other tasks, those expenses have become an increasingly important consideration for technology companies investing heavily in AI infrastructure.

"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.

AI Spending Becoming a Resource Management Issue

Mosseri compared AI token budgets with other operating expenses that organisations already allocate across teams, including payroll, computing hardware, storage, and labelling budgets. He said AI usage could eventually be managed using similar principles, with companies distributing available resources based on business priorities.

"I think of it like...any other resource," Mosseri 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."

According to Mosseri, any future spending limits would likely reflect how effectively engineers use AI tools. He said budgets could be tied to whether an employee demonstrates a positive return on investment from their AI usage rather than applying identical limits across the company.

Although Meta does not currently impose AI token caps on employees, Mosseri said he believes they could become a practical approach as enterprise AI adoption expands. He also expects AI token costs to decline over time if model providers compete more aggressively on pricing.

Industry-Wide Focus on Rising AI Costs

Meta is not the only technology company reassessing AI spending. The company recently shut down an internal AI token-spending leaderboard after costs reportedly put it on track to reach billions of dollars in AI spending in 2026.

Other companies have also adjusted their AI strategies. Uber reportedly exceeded its planned 2026 AI coding budget by April, while Microsoft discontinued some Claude Code licenses and consolidated engineers around its Copilot CLI tool as organisations sought to manage AI-related operating costs better.

Mosseri said Meta has already reduced some unnecessary spending by eliminating what he described as “silly things,” including the internal token leaderboard. He added that simply consuming more AI tokens does not necessarily produce greater business value.

"It's not that hard to build a token incinerator, and that doesn't create a lot of value," Mosseri said.

As enterprises continue integrating generative AI into software development, companies are increasingly balancing broader access to AI tools with the financial realities of operating large-scale AI infrastructure.