Uber has capped employee spending on artificial intelligence (AI) coding tools at 1,500 dollars a month per tool, after burning through its annual AI budget in just four months.
The limit applies separately to each agentic coding tool, including Anthropic’s Claude Code and Cursor, a company spokesperson told Bloomberg. Employees can track their usage on an internal dashboard and request approval to exceed the cap in certain cases.
The move reverses a more permissive approach. Uber had earlier urged engineers to use the tools as widely as possible and reportedly ranked their usage on internal leaderboards.
The company’s Chief Technology Officer, Praveen Neppalli Naga, disclosed in April that Uber had exhausted its full year AI budget within four months.
Uber President and Chief Operating Officer Andrew Macdonald has since questioned the payoff, saying it was “very hard to draw a line” between rising AI use and the delivery of new consumer features.
Chief Executive Dara Khosrowshahi said this month that about 10 percent of Uber’s code is now written by autonomous AI agents, underlining how deeply the company has embedded the technology.
Uber’s experience mirrors a wider corporate dilemma. Firms are pouring money into AI tools whose token based costs can climb sharply the more they are used, while struggling to show clear productivity returns.
Uber described the caps as a way to encourage AI adoption responsibly and has not signalled any retreat from its broader AI strategy.
