May 26, 2026
Surge pricing, but for AI
Uber blows through its AI budget in 1 quarter
Uber pushed workers to use more AI, then seemed shocked when the bill exploded
TLDR: Uber reportedly blew through a full year’s AI coding budget in four months after pushing teams to use the tools more, even though executives can’t clearly show riders got much benefit. Commenters called it predictable management chaos, with jokes about “breeding cobras” and warnings that this looks like another expensive tech hype cycle.
Uber wanted an AI gold rush and got an AI bar tab instead. The company reportedly burned through its 2026 budget for AI coding tools in just four months, after using an internal leaderboard that ranked teams by how much they used the tools. And the internet’s reaction? Somewhere between “well, obviously” and full popcorn-munching disbelief. One commenter summed up the mood perfectly with a snake-infested management parable: if you pay by the dead cobra, don’t act surprised when people start breeding cobras. Ouch.
That line became the unofficial theme of the discussion. Critics said Uber basically turned AI usage into a contest, then acted stunned when employees ran up the meter. Others were less angry and more amused, arguing the incentives actually worked too well: if leadership told everyone to use as much AI as possible, what exactly did they expect? There was also a more serious undercurrent: several people saw this as a classic Silicon Valley repeat of the dot-com era, when companies threw money at trendy tools before proving they were actually useful. As one commenter put it, this could be the moment big companies realize AI hype and real-world payoff are not the same thing.
The bigger twist is that even Uber’s own executive admitted it’s hard to prove all this spending has led to more useful features for riders. That’s why this story hit a nerve: it’s not just about one company overspending, it’s about whether corporate AI fever is becoming the new office treadmill—expensive, impressive-looking, and maybe not taking anyone anywhere. Meanwhile, some commenters asked the practical question hanging over the whole mess: couldn’t Uber just switch to a cheaper option?
Key Points
- •Andrew Macdonald said Uber cannot yet clearly connect higher Claude Code usage to a proportional increase in useful consumer-facing features.
- •The article reports Uber exhausted its 2026 AI coding tools budget in four months after encouraging employee adoption through an internal usage leaderboard.
- •Dara Khosrowshahi said about 10% of Uber’s committed code is built by autonomous agents, with AI tools also used across legal, marketing, and engineering teams.
- •Gartner research cited in the article says advanced-model inference costs may fall sharply by 2030, but enterprise AI costs may remain high because agentic models use more tokens and providers may not pass through all savings.
- •Uber spent $951 million on research and development in the first quarter of 2026, up nearly 17% year over year, while continuing to invest in autonomous driving.