Uber’s aggressive push into artificial intelligence is running into an uncomfortable question inside the company: is all that AI spending actually paying off?
In April, internal alarm bells rang after Uber’s CTO Praveen Neppalli Naga revealed the company had already blown through its 2026 budget for Claude Code, a popular AI coding assistant, just a few months into the year.1 The remark triggered what operations chief Andrew Macdonald later described as a “head-exploding moment,” forcing leaders to confront how rapidly AI token costs were mounting and what trade-offs that implied for other expenses, including hiring.1
By May, Macdonald was publicly questioning the return on that investment. In a Rapid Response interview, he said it was becoming “harder to justify” the money spent on AI, noting that rising token consumption was not matched by a proportional rise in useful features for riders and drivers.1 “That link is not there yet,” he said, adding that it was “very hard to draw a line” from AI usage metrics to “25 percent more useful consumer features.”1
A separate account of the interview underscored the same concern: Uber “isn’t seeing a connection between rising token consumption for Claude Code and more useful features being delivered to consumers,” even as internal metrics around AI usage surge “in a really astronomical direction.”2
Uber’s leadership has tried to offset AI costs by slowing the growth of its workforce. CEO Dara Khosrowshahi told investors earlier this month that the company was hiring fewer people to counter rising AI investments, effectively pitting “token consumption and the associated cost versus headcount.”2
Macdonald stressed that, for casual users, AI can appear “free” — but for companies footing the infrastructure bill, the calculus is changing.1 Uber spent $3.4 billion on research and development in 2025, up 9 percent year over year, and Macdonald now warns that, until there is a clear link between AI usage and concrete product gains, such trade-offs “become harder to justify.”2