Uber blew through its entire annual AI budget in four months. The company rolled out Anthropic’s Claude Code to roughly 5,000 engineers in December 2025, and by April 2026, the money earmarked for AI tools like Claude Code and Cursor was gone.
The kicker: leadership isn’t even sure it was money well spent.
The numbers are staggering, the results are not
By spring 2026, 95% of engineers were using AI tools on a monthly basis. Around 70% of code commits were AI-driven. Usage of agentic AI features surged from 32% in February to 84% by March 2026.
Uber COO Andrew Macdonald put it bluntly in a May 2026 interview with Rapid Response.
“That link is not there yet, right? I think maybe implicitly there is more that is getting shipped, but 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.'”
Monthly API costs per engineer ranged from $500 to $2,000. Uber’s total R&D expenditure hit $3.4 billion in 2025, already a 9% jump year-over-year. The 2026 AI budget was supposed to last twelve months. It lasted four.
The internal reckoning
CTO Praveen Neppalli Naga’s disclosures in April 2026 reportedly triggered an internal review of AI spending at the company. The question on the table was uncomfortable but necessary: should Uber keep escalating AI investments, or should it start reassessing headcount in light of these costs?
One particularly striking data point: 11% of live backend updates were being executed by AI agents with zero human oversight.
Uber isn’t alone in hitting this wall. Microsoft has reportedly placed restrictions on Claude usage due to escalating expenditure.
What this means for investors watching AI and crypto
The core problem Uber just exposed is that variable token pricing creates budgeting nightmares for enterprises. That dynamic has direct implications for demand around AI-related computational capacity, which is exactly where several crypto-native projects position themselves.
Decentralized compute networks like Akash, Render, and io.net have pitched themselves as cheaper alternatives to centralized AI infrastructure providers. The crypto projects most exposed are those whose valuations are implicitly tied to the assumption that enterprise AI demand will grow exponentially and indefinitely. Uber demonstrated that demand can grow exponentially while budgets collapse on a much shorter timeline.