Meta Considers Capping AI Token Budgets Per Engineer
Meta Platforms is exploring limits on AI token consumption for individual engineers. According to Adam Mosseri, head of Instagram, this measure may become necessary in the near future.

Meta Platforms, the parent company of Instagram, is considering managing its artificial intelligence expenses by capping the digital token usage for each engineer. Adam Mosseri, head of Instagram, recently presented projections indicating that this limitation could become relevant within the next one to two years.
Mosseri compared AI token costs to other essential business expenditures like salaries or operational expenses. He suggested in a podcast that the AI token consumption of a strong engineer might eventually rival their daily salary. In such a scenario, setting consumption caps would likely be required to control costs.
The management of AI token expenses has emerged as a key topic of discussion within technology companies. Meta itself shut down an internal AI token spending leaderboard after costs threatened to reach billions of dollars in 2026, aiming to balance innovation with efficient resource allocation.
Meta is not the only tech company re-evaluating its AI usage and costs. Uber also faced a similar situation when it exceeded its AI coding budget within four months. Microsoft, on the other hand, has had to reduce certain external AI licenses to focus on its own internal tools.