Meta executive says AI spending may soon be capped per engineer

In short: Meta’s Adam Mosseri says companies may soon set per-engineer limits on how much employees can spend using AI tools.

What's going on

Adam Mosseri, who runs Instagram at Meta, said he can imagine a near future where companies have to control “AI token” spending the way they control other business costs. He said this could happen within a year or two.

Tokens are the small units used to measure how much text an AI system processes. Think of it like a meter on a taxi ride. The longer the conversation and the bigger the requests, the higher the bill.

Mosseri said a strong engineer’s AI usage could eventually cost as much as that person’s salary. In that situation, he expects companies to introduce caps, meaning a limit on how many tokens an employee can spend.

Meta does not currently have token caps, according to Mosseri. He also said Meta reduced some waste by shutting down “silly” internal projects, including a token spending leaderboard that showed who used the most.

Other companies are also tightening controls on AI costs. TechCrunch noted that Uber capped employee AI spending after using up its 2026 budget early, and Microsoft reportedly canceled some third-party AI coding tool licenses and pushed engineers toward its own tools.

What to watch

If more employers start treating AI usage like an expense account, everyday AI features could be shaped by budgets. Teams may need approval for heavy AI use, and they may pick cheaper tools or smaller models. Mosseri also said token costs could fall later if AI providers compete on price.

Source: TechCrunch AI