$500 to $2,000 per engineer, per month. That's what Claude Code is costing Uber, and the company burned through its entire 2026 AI coding budget by April — four months into the year.
Uber rolled out Claude Code to engineers in December 2025. Adoption moved faster than anyone at the company anticipated. By April, 95% of Uber's engineers were using AI tools at least once a month, and 70% of all committed code had AI involvement. The company's CTO has said Uber is "back to the drawing board" on how to budget for AI tooling.
What $2,000/Month Per Engineer Actually Means
At the low end, $500/month is manageable for a company Uber's size. At the high end, $2,000/month per engineer is a different problem entirely. Multiply that across thousands of engineers and you're looking at costs that can breach $100 million annually — just for AI coding assistance.
The issue isn't that Claude Code failed. It's that it worked well enough for engineers to use it constantly. Most enterprise software budgets assume partial adoption: you buy 1,000 licenses and 600 people use them occasionally. AI coding tools with token-based pricing — where you pay based on how much text the model reads and writes, not a flat monthly fee — don't behave that way. Heavy users consume dramatically more than light users, and the engineers most committed to AI-assisted development are almost always the heaviest users.
The Trap Every Company Is About to Fall Into
Uber isn't a special case. They're the first large tech company to talk publicly about blowing past AI budget projections, but the same math is running at every company that handed out AI coding tool access in late 2025 and assumed adoption would plateau at moderate levels.
The $500-$2,000 range also reveals how uneven consumption can be. An engineer using Claude Code for occasional code review might spend $50/month. An engineer doing full AI-assisted feature development all day might spend $2,000. Flat-rate enterprise contracts make this more predictable, but usage-based pricing — which many deployments default to — punishes successful adoption.
For anyone currently planning an AI coding rollout, Uber's experience is worth internalizing. A 50-engineer pilot doesn't reveal what happens when 5,000 engineers have access and actually like the tool. Budget for the high-adoption scenario. If usage stays low, the money sits unspent. If it takes off the way it did at Uber, you won't be explaining to finance in April why the annual AI budget ran out before summer.