Meta is stepping back from "tokenmaxxing," a trend in AI usage, as the company moves to curb employee AI costs. According to an internal memo, Meta plans to impose token limits on employee AI usage, as internal AI usage is projected to cost Meta billions in 2026. The company is building an internal platform called AI Gateway to monitor usage and spending in real time, set budgets, and establish limits on employees' token spend. The memo, shared with about 6,000 staffers, stated: "We’ve seen an exponential increase in AI usage and [we] are tracking to spend billions on internal use alone in 2026." For 2027, Meta expects to manage AI tokens with budgets, allocation decisions, and supporting tools. As part of the spending crackdown, a team has created a central dashboard called AI Gateway to monitor usage and spending, with automated alerts for unusual spending spikes. Meta is also preparing to encourage staff to shift from third-party AI tools toward in-house solutions like its coding assistant, MetaCode (previously called Devmate), while allowing continued access to new third-party models.