Uber's AI Paradox: Cutting Customer Service Jobs While Burning Through Its AI Budget

Uber said the customer service operation had become "too complex and siloed," and the cuts were part of a broader effort to "embrace AI," simplify its structure, and improve efficiency . A company spokesperson said the move aimed at simplifying processes with AI handling more of the workload . As part of the restructuring, remote employees were told they would need to relocate to designated hub offices .

Contrast with the June 2026 Layoffs

Just seven weeks earlier, on June 3, 2026, Uber had cut 23% of its "People and Places" division — the team handling HR, recruitment, talent acquisition, workplace facilities, and culture . Those cuts disproportionately hit senior roles and represented less than 1% of Uber's ~34,000 global employees .

The critical difference: Uber explicitly did not attribute the June cuts to AI. The restructuring was led by newly promoted president Jill Hazelbaker to "simplify team structure" . CEO Dara Khosrowshahi said the changes were "essential to optimize the productivity of the People team" . Some reports carried headlines like "Uber slashes 23% of HR and recruitment jobs, says AI isn't the reason" .

In June, Uber was at pains to say AI was not the cause. In July, it was the first time they explicitly said AI was the cause .

The CEO's Earlier Signals

At the World Economic Forum in Davos in January 2026, Khosrowshahi sat for a WSJ interview titled "Why Uber Is Letting AI Use 'Common Sense' in Customer Service." He explained that Uber was scrapping rigid rule-based systems and empowering AI to make judgment calls — using context and "common sense" to deliver better customer service experiences rather than sticking to predefined scripts .

Khosrowshahi was also blunt about companies engaging in superficial AI adoption, calling out firms that are "play-acting" their way into a pretend transformation rather than pursuing genuine AI transformation . The January Davos interview laid the philosophical groundwork: Uber was already moving from human-driven, rule-based customer service to AI-driven judgment. The July layoffs were the operational consequence of that strategic shift.

The AI Budget Blowout

Meanwhile, Uber's AI spending was spiraling out of control. The company completely exhausted its entire 2026 internal AI tooling budget by the end of April 2026 — just four months into the fiscal year . The budget was sized for ~12 months of use, but the real burn rate ran at approximately 3x the projection .

What drove the burn? Claude Code (Anthropic's AI coding assistant) saw massive organic adoption across ~5,000 Uber engineers . An internal leaderboard ranking AI usage volume drove Claude Code adoption from 32% to 84% of engineers . Per-engineer costs ranged from $500 to $2,000 per month .

In early June, Khosrowshahi stated, "We exceeded our AI budget for the entire year in just one quarter," and confirmed the high spending had forced the company to slow hiring .

Usage Caps Imposed

On June 2, 2026, Bloomberg reported Uber instituted a $1,500 monthly token spending cap per employee per AI coding tool — meaning an engineer running both Claude Code and Cursor gets $1,500 for each . The caps were a direct cost-control measure after the budget blowout . Multiple reports confirm that AI tools now generate at least 10% of Uber's production code , underscoring that AI integration is not a side experiment but a core part of engineering output.

The Broader Industry Tension

The pattern reflected here is industry-wide. The Economic Times summarized it as "AI becomes latest driver of tech layoffs" and noted the same dynamic at Amazon, Microsoft, and other Big Tech firms . Amazon also announced job cuts on the same day as Uber's July 23 announcement .

The tension is stark: Uber blew through its entire annual AI budget in four months on tools like Claude Code, then capped employee AI usage to control costs. It then laid off 10% of customer service workers explicitly citing AI as the reason — just seven weeks after a separate 23% cut to its People and Places division that was not AI-related .

The contradiction — companies spending aggressively on AI while using that same AI to justify headcount reductions — is a central theme of the 2026 tech labor market.