On June 2, 2026, Bloomberg reported that Uber has put a monthly cap of $1,500 per employee on each agentic coding tool — Claude Code and Cursor both fall under the rule. Usage is trackable on an internal dashboard, and in certain cases the cap can be exceeded with permission.
The backstory matters more: back in April, Uber’s CTO admitted the company had burned through its entire annual AI budget in just four months. Going from “use AI as much as possible” to monthly rationing makes Uber one of the first large US tech companies to formally cap employee AI usage — and puts enterprise AI cost governance squarely on the table.
From “Use Everything” to Rationing
The Information reported earlier that Uber had encouraged staff to use AI “as much as possible” and even ranked employees on internal leaderboards by AI usage. The turn came in April, when the CTO showed internally that tools like Claude Code can blow through AI budgets — the full-year budget was gone in a third of the year. The June 2 policy is a formal admission that uncontrolled agentic tool usage now costs enough, per person, to require limits. It also quietly retires the leaderboard logic: if usage was a virtue signal in Q1, by mid-year it had become a line item someone has to defend.
Note the mechanics: a cap, a dashboard, and an exception process — not a tool ban. Uber keeps the flexibility while giving every employee a first clear look at their own AI bill.
What $1,500 a Month Buys
Agentic coding tools bill by token consumption, which is nothing like a flat per-seat subscription. An engineer leaning hard on Claude Code or Cursor — agents reading and rewriting code, calling tools, retrying failed steps — can easily push a monthly bill into four figures. That is exactly why GitHub’s move to token-based Copilot billing in late May drew howls from developers: the pricing model itself is still violently in flux, while finance departments are asked to budget for it a year at a time.
A $1,500 ceiling is a pragmatic middle ground: enough for most day-to-day engineering work, but a hard stop on runaway agent loops and the “let it run overnight” style of waste. It also forces a conversation the industry has been avoiding — what a reasonable monthly AI bill per engineer actually looks like.
The ROI Problem
The genuinely hard part is not saving money — it is measuring value. COO Andrew Macdonald said on a podcast in May that it is “very hard to draw a line” between AI usage and new consumer features actually shipping. A Bain survey cited by Bloomberg on June 1 found AI has delivered less cost reduction than most firms predicted, and the Wall Street Journal reports Corporate America is starting to ration AI as costs skyrocket.
Uber’s predicament is everyone’s predicament: spend is immediate and measurable, while returns are smeared across productivity gains nobody can cleanly attribute. Blowing the annual budget in four months just detonated the structural problem early. The cap is a tourniquet, not an answer — the answer is treating AI spend as an engineering resource with unit costs, not an employee perk. Until teams can say what a shipped feature cost in tokens, budgets will keep getting set by panic instead of by planning.
What It Means for Engineering Teams
Three practical lessons. First, usage monitoring for agentic tools should be built in from day one, not bolted on after the budget explodes — dashboards, unit costs, and anomaly alerts belong in the default setup now. Second, pricing is shifting from per-seat to consumption-based; procurement contracts and cost models need redesign, and token-level invoices will keep getting more common. Third, what you should measure is tasks completed and rework rate, not raw token counts — the same theme that runs through our lessons from Cursor’s cloud agents: the more agents you run, the more governance and feedback loops matter.
Sources
- Uber caps usage of AI tools like Claude Code to cut costs — Bloomberg
- Uber caps employee AI spending after blowing through budget in 4 months — TechCrunch
- Uber CTO shows Claude Code can blow AI budgets — The Information
AI-assisted summary compiled from the sources above, reviewed by a human before publishing.