Meta caps employee AI token usage after internal AI spending forecast reaches billions for 2026
Meta $META is capping how many AI tokens its own staff can use after internal AI spending forecasts hit billions of dollars for 2026. A new system called AI Gateway will track usage and flag spikes, and employees are being pushed toward Meta's in-house MetaCode assistant.
The cost of running AI internally has grown large enough that even Meta $META is putting its own employees on a budget. According to a staff memo, the company is moving to limit how many AI tokens employees can consume after internal forecasts put its in-house AI spending in the billions of dollars for 2026. To get the spending under control, Meta is rolling out a platform called AI Gateway that tracks AI usage and cost across teams in one place, with automated alerts to flag unusual spikes. Employees are being steered toward MetaCode, Meta's own coding assistant, to reduce reliance on more expensive third-party tools. The company expects to manage tokens through a more structured framework by 2027, including formal budgets and allocation decisions. The move is a reminder that generative AI is not just a capital-expenditure story about data centers and chips. The per-query cost of giving thousands of employees access to powerful models adds up quickly, and companies that encouraged broad internal adoption are now confronting the bill. Meta's response, building internal tooling and metering consumption, is the kind of cost discipline other large employers are likely watching closely.