Meta sent an internal memo to 6,000 employees warning that internal AI usage costs are approaching billions of dollars in 2026, driven by an exponential increase in token consumption. Employees consumed 73.7 trillion tokens in roughly 30 days, tracked on an internal leaderboard called "Claudeonomics" that inadvertently incentivized volume over productivity. CTO Andrew Bosworth pushed back on "tokenmaxxing" in a separate memo, stating "All motion is not progress and token usage alone is not a measure of impact of any kind." The company plans to replace the leaderboard with a centralized "AI Gateway" dashboard.

The spending surge occurs as Meta commits $135 billion to AI infrastructure through 2026 and $600 billion to data center buildouts through 2028. Unlike capital expenditures on GPUs, internal token bills represent a less-discussed cost layer from third-party AI tools. Uber exhausted its entire 2026 AI coding budget in four months, capping employee spending at $1,500 per month per tool. Both companies reveal the gap between aggressive AI adoption and measurable productivity gains.

A KPMG survey found only 26% of companies have comprehensive AI cost visibility, enabling unchecked spending growth. Goldman Sachs projects a 24x rise in enterprise token consumption by 2030, reaching 120 quadrillion tokens monthly. OpenAI CEO Sam Altman acknowledged that companies report spending heavily while knowing there is significant waste.

Meta will deploy the AI Gateway dashboard in coming weeks, with full budget frameworks operational by early 2027. The shift from open-ended AI access to metered usage represents a cultural change at a company that has aggressively promoted internal AI adoption. This cost governance approach may slow integration pace or improve spending efficiency, directly impacting Meta's competitive AI advantage narrative.