Uber blew through its entire 2026 AI budget by mid-April. Four months into the year, the well was dry, and the company’s top executives are now publicly asking a question that a lot of tech leaders are thinking privately: is any of this actually working?
Uber President and COO Andrew Macdonald put it bluntly in a recent interview. The company can’t draw a clear line between skyrocketing AI token consumption and the delivery of meaningfully better features to consumers. Adoption metrics look incredible on paper, but the real-world output? Fuzzy at best.
The numbers tell a strange story
Here’s what happened. Uber rolled out Anthropic’s Claude Code to roughly 5,000 engineers back in December 2025. Adoption was volcanic. Usage of agentic coding features surged from 32% in February to 84% by March 2026. By that point, 95% of Uber’s engineers were using AI tools on a monthly basis, and nearly 70% of code commits involved them in some capacity.
Macdonald said in a May 2026 interview that the connection between rising token consumption and useful consumer features simply isn’t there yet. “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.
