Uber's aggressive push into AI coding tools has hit a financial wall, with the company exhausting its planned AI budget just months into 2026. Despite spending $3.4 billion on research and development in 2025, Chief Technology Officer Praveen Neppalli Naga revealed that rapid adoption of Anthropic's Claude Code has driven costs far beyond internal projections. Engineers were encouraged to use AI tools, with internal leaderboards tracking usage.

Claude Code quickly became the dominant tool, with usage surging since late 2025 while Cursor usage plateaued. The company is now testing OpenAI's Codex as it expands its AI stack. Uber's R&D expenses rose 9% to $3.4 billion in 2025, with the company expecting this figure to keep climbing as AI adoption accelerates. This suggests AI may be as much a cost driver as a productivity lever.

The payoff is showing: around 11% of Uber's live backend code updates are now written by AI agents, up sharply in recent months. These systems power ride-matching, pricing, and bug fixes. Naga envisions a future with "agent engineers"—AI systems that handle coding, testing, and deployment autonomously, supervised by other AI tools. As AI takes on more workload, Uber's aggressive push raises questions about the long-term role of human engineers in the company's development process.