Meta Caps Internal AI Token Spending After Costs Approach Billions in 2026

Meta Caps Internal AI Token Spending After Costs Approach Billions in 2026 - Meta sent an internal memo to 6,000 employees warning that internal AI usage costs are approaching billions of dollars in 2026 [1] - Employees consumed 73.7 trillion tokens in roughly 30 days, tracked on an internal leaderboard called "Claudeonomics" [2] - CTO Andrew Bosworth said in a separate memo: "All motion is not progress and token usage alone is not a measure of impact of any kind" [1] - Meta will deploy a centralized "AI Gateway" dashboard and implement formal token budgets starting in 2027 [2] - The company is steering employees away from Anthropic's Claude toward its own MetaCode coding assistant [1] Meta is imposing centralized spending controls on employee AI usage after internal token consumption surged to levels that put the company on track for billions of dollars in costs during 2026, according to an internal memo first reported by The Information [1]. The memo, sent to approximately 6,000 employees, flagged an "exponential increase" in AI usage and warned that teams currently have limited visibility into their own consumption. CTO Andrew Bosworth followed with a separate memo pushing back on what the company internally dubbed "tokenmaxxing" — the practice of employees inflating AI usage metrics, sometimes through gamified leaderboards rather than genuine productivity gains. "Nobody should be using AI tools just for the sake of using them," Bosworth wrote. "All motion is not progress and token usage alone is not a measure of impact of any kind" [1]. The crackdown comes as Meta plans to spend up to $135 billion on AI infrastructure through 2026 and commits $600 billion to data center buildouts through 2028 [3]. The internal consumption problem adds a new, less-discussed cost layer: the token bills employees rack up using third-party AI tools for day-to-day coding and productivity work. What the Memo Says The internal memo disclosed that Meta employees consumed 73.7 trillion tokens in roughly 30 days, a figure tracked on an internal leaderboard called "Claudeonomics" — a reference to Anthropic's Claude, one of the third-party AI tools widely used inside the company [2]. The leaderboard, which ranked employees and teams by token consumption, inadvertently incentivized usage volume over productive output. Meta plans to dismantle the leaderboard and replace it with a centralized monitoring platform called "AI Gateway," which will track usage and spending across teams in real time [2]. The system will include automated alerts for unusual spending spikes. Full token budgets and formal allocations will take effect in 2027 [1]. The company is also steering employees toward MetaCode, its proprietary coding assistant, and away from external tools like Anthropic's Claude [1]. The shift serves a dual purpose: reducing third-party API costs and dogfooding Meta's own AI products. Why It Matters Meta's token spending problem is not unique. Uber exhausted its entire 2026 AI coding budget in four months, prompting it to cap employee spending at $1,500 per month per tool [4]. At Uber, nearly 95% of engineers use AI tools monthly and close to 70% of committed code is AI-generated, yet COO Andrew Macdonald said the link between token spending and measurable output "is not there yet" [4]. A KPMG survey found that only 26% of companies have comprehensive visibility into their AI costs [3]. The opacity has allowed spending to balloon unchecked. Goldman Sachs projects a 24x rise in enterprise token consumption by 2030, reaching 120 quadrillion tokens per month across the industry [3]. OpenAI CEO Sam Altman has acknowledged the dynamic, noting that companies are telling him "I am spending a ton of money on AI...but I know there's a ton of waste" [3]. The pattern suggests a broader industry reckoning as enterprises move from aggressive AI adoption to cost governance. Market Context Meta shares traded at approximately $567 on June 13, giving the company a market capitalization of roughly $1.4 trillion [5]. The stock has pulled back from a 52-week high of $796.25, reflecting broader investor scrutiny of Big Tech AI spending levels [5]. The internal cost pressures add nuance to Meta's AI investment narrative. While CEO Mark Zuckerberg has publicly committed to massive infrastructure buildouts, the memo suggests that even internal consumption — distinct from the capital expenditure on GPUs and data centers — is becoming a material cost center that requires active management. What's Next Meta's AI Gateway dashboard is expected to deploy in coming weeks, with the full budget framework operational by early 2027 [2]. The transition from open-ended AI tool access to metered usage marks a significant cultural shift at a company that has aggressively encouraged AI adoption across its workforce. The move positions Meta alongside Uber and other large employers that are discovering the gap between AI adoption metrics and actual productivity gains. For investors, the question is whether token governance translates into more efficient AI spending — or whether it simply slows the pace of internal AI integration that Meta has pitched as a competitive advantage.