Uber has rolled out strict internal spending limits on AI tools after the ridesharing company burned through its entire annual artificial intelligence budget in just four months, according to a Bloomberg report. The new policy places a $1,500 monthly cap per employee for each agentic coding tool they use, including Anthropic's Claude Code and Cursor. Uber has built an internal dashboard so workers can monitor their own consumption, though managers can grant exceptions when projects require heavier usage.
The clampdown comes a few months after Uber's CTO disclosed in April that the company had already exhausted its full-year AI budget by spring. The runaway spending reportedly followed an aggressive internal campaign that urged employees to use AI "as much as possible," with the company even ranking staff on competitive leaderboards based on their usage, according to earlier reporting by The Information. The push appears to have backfired, driving costs well beyond what leadership anticipated.
The episode highlights growing skepticism inside Uber about what AI is actually delivering. COO Andrew Macdonald recently suggested on a podcast that it's "very hard to draw a line" between AI usage and tangible new consumer features, casting doubt on the productivity gains that have justified massive enterprise spending across the tech industry. His comments echo a broader concern: as companies pour billions into AI tools and infrastructure, concrete returns on investment have remained largely theoretical, leaving some executives increasingly impatient as they wait for the payoff to materialize.