Meta's Mosseri: AI token budgets may be capped per engineer within 1-2 years
Original: Meta’s Adam Mosseri says AI token budgets could soon be capped per engineer
Why This Matters
Rising AI token costs are forcing major tech companies to rethink resource allocation and engineer spending policies.
Instagram head Adam Mosseri stated on Lenny's Podcast that Meta may need to cap AI token spend per engineer within one to two years, as token burn rates for a strong engineer could soon equal their salary or cost of employment.
In an appearance on Lenny's Podcast, Meta's Instagram head Adam Mosseri said per-engineer AI token caps could become necessary within one to two years. He stated: '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 compared token budgets to existing resource management decisions — GPUs, storage, OpEx, and headcount — arguing they must be managed the same way, with caps proportional to each engineer's ability to deploy them in an 'ROI-positive' manner. Meta currently has no token caps in place, but the company previously shut down an internal AI token spend leaderboard after costs put it on track for billions of dollars in 2026. Mosseri noted the company has curbed costs by eliminating 'silly things' like that leaderboard, saying, 'It's not that hard to build a token incinerator, and that doesn't create a lot of value.' Meta is not alone: Uber exhausted its 2026 AI coding budget by April, and Microsoft canceled Claude Code licenses to consolidate engineers around its Copilot CLI. Mosseri also expects AI token prices to fall long-term as model providers compete for users.