Uber's AI Push Hits a Wall–CTO Says Budget Struggles Despite $3.4B Spend

Uber raced into AI—and tripped over the bill. After a $3.4B research spend last year, the CTO says Uber burned through its 2026 AI budget early, largely thanks to a company-wide push to use coding bots like Anthropic’s Claude and Cursor. Leaderboards encouraged heavy use, and surprise: heavy use costs money. Now they’ll test OpenAI’s Codex, while AI already writes about 11% of live code updates. The community? Spicy.

The top reaction is pure skepticism about ROI. As one put it, “If AI boosts productivity, why isn’t it paying for itself?” Others dunked on quality, pointing to cringe Uber Eats blurbs—like the now-memed “Customisable burgers with 250,000+ toppings”—as proof the bots can be tone-deaf. Cue jokes about AI interns swiping the corporate card and “a vibrant culinary journey” haunting the app.

Then came the headline police. Several argue the story’s framing is off. “Holy misleading headlines Batman,” one wrote, insisting the $3.4B is overall R&D (salaries, hardware, everything), not just AI tokens; the real issue is usage exploding faster than finance planned. Another demanded a receipt-level breakdown—“Are they sending $1B to Anthropic?”—calling the piece vague. So the split is set: Is AI a money pit or a rocket booster? Uber’s betting on “agent engineers,” but the crowd wants proof it won’t surge-price the budget forever.