Uber imposes $1,500 monthly cap on employee AI spending after annual budget exhausted

Uber has introduced internal monthly limits on AI spending after its chief technology officer disclosed in April that the company had burned through its entire annual AI budget in roughly four months. The new rule sets a $1,500 cap per employee and per agentic coding tool and is intended to rein in costs while preserving oversight. According to reporting cited by the article, the cap explicitly covers agentic coding tools such as Anthropic’s Claude Code and Cursor. Employee usage of those tools is visible through an internal dashboard, and staff can exceed the monthly caps only with company permission.

The policy shift follows earlier internal signals that encouraged heavy AI use: employees were reportedly urged to use AI “as much as possible” and saw their usage ranked on internal leaderboards. Uber’s chief operating officer, Andrew Macdonald, has questioned the measurable productivity gains from AI, saying “it’s very hard to draw a line” between AI usage and new consumer features.

The cap reflects a broader industry reckoning with AI spending and uncertain return on investment, as some companies pull back to moderate costs while awaiting clearer ROI. Allowing exception requests suggests Uber is trying to balance cost control with continued experimentation, but the change underscores growing internal and external scrutiny of enterprise AI expenditures.