Uber blew through its entire 2026 AI budget by mid-April. Four months into the year, the funds were exhausted, and top executives are now publicly questioning whether the spending is producing tangible results.
Uber President and COO Andrew Macdonald said the company cannot draw a clear link between rising AI token consumption and the delivery of better consumer features. Adoption metrics look strong on paper, but the real-world output remains unclear.
The Numbers Tell a Strange Story
Uber rolled out Anthropic's Claude Code to roughly 5,000 engineers in December 2025. Usage of agentic coding features surged from 32% in February to 84% by March 2026. By that point, 95% of Uber's engineers used AI tools monthly, and nearly 70% of code commits involved them.
Macdonald said the connection between rising token consumption and useful consumer features is not yet visible. "It's very hard to draw a line between one of those stats and, 'Okay, now we're actually producing 25 percent more useful consumer features,'" he noted.
Meanwhile, per-engineer monthly API costs ballooned to between $500 and $2,000, blowing past internal forecasts. CTO Praveen Neppalli Naga confirmed the full annual AI budget was exhausted by mid-April. Uber's total R&D spending hit $3.4 billion in 2025, a 9% increase year-over-year.
What Investors Should Watch
Per-engineer API costs of $500 to $2,000 per month are steep, suggesting AI inference providers like Anthropic hold substantial pricing power. Unsustainable costs tend to trigger one of two responses: companies find cheaper alternatives, or they simply use less.
The broader lesson: adoption is not the same as value creation. Usage metrics are not productivity gains. Uber's candid admission that the link between spending and results "is not there yet" should land as a meaningful signal.