Uber Caps Employee AI Tool Spending at $1,500 Per Month After Budget Crisis

Quick Facts

  • Uber burned through its entire 2026 AI coding budget in just four months due to soaring token consumption
  • Individual engineers were generating monthly bills of $500-$2,000 using AI coding tools like Claude Code
  • Company now limits spending to $1,500 per month per tool for each employee

Uber Technologies implemented strict spending caps on employee AI tool usage after exhausting its entire planned 2026 artificial intelligence budget in just four months.

The ride-hailing giant now limits individual employees to $1,500 per month per AI coding tool, according to CTO Praveen Neppalli Naga. The caps apply separately to each tool, meaning spending on Cursor does not affect an employee’s Claude Code budget.

Before the restrictions, software engineers were generating monthly bills ranging from $500 to $2,000 in token consumption. Naga himself spent $1,200 during a two-hour personal demonstration of Claude Code.

“I’m back to the drawing board because the budget I thought I would need is blown away already,” Naga told The Information in April.

The financial strain reflects rapid adoption across Uber’s engineering teams. By March 2026, 84% of Uber’s developers were using agentic coding tools, and AI agents now write approximately 11% of the company’s live backend code updates.

COO Andrew Macdonald called the CTO’s budget disclosure a “head-exploding moment” that triggered company-wide discussions about token consumption costs. He noted the company has not yet established a clear connection between AI tool usage and consumer product improvements.

Uber deployed an internal dashboard allowing employees to track real-time AI activity and costs. The company also created a formal review process for engineers to request permission to exceed standard limits for heavy workloads.

The budget crisis contributed to a 3.1% decline in Uber’s stock price on Tuesday. The company’s market capitalization stands at approximately $146.4 billion.

Uber’s experience reflects broader enterprise struggles with AI costs. Gartner analysis shows agentic AI models require 5-30 times more tokens per task than standard chatbots. The average enterprise AI budget has grown from $1.2 million in 2024 to $7 million in 2026.

Other companies face similar challenges. Microsoft pulled back Claude Code licenses after per-engineer monthly bills climbed to $500-$2,000. One consulting client reportedly spent half a billion dollars in a single month on Claude usage.

“We think this is all a pretty straightforward way to responsibly encourage agentic AI adoption and experimentation at scale across the company,” an Uber spokesperson said.

Read more: Uber caps employee AI spending after blowing through budget in 4 months

Get updates

Get curated daily technology news in your inbox.

Discover more from The SaaS Sentinel

Subscribe now to keep reading and get access to the full archive.

Continue reading