TLDRs:
- Uber says AI tools now produce about 10% of internal codebase
- Company introduces spending caps after exhausting its AI coding budget
- Developers rely on tools like Cursor and Claude Code
- Broader industry faces rising costs and tightening AI software budgets
Uber has revealed that artificial intelligence is now responsible for generating roughly 10% of its internal code, marking a significant shift in how one of the world’s largest ride-hailing and delivery platforms builds and maintains its software systems. The milestone underscores how quickly AI-powered development tools have moved from experimental use cases into core engineering workflows.
According to senior executives, the company’s AI adoption has accelerated across multiple departments, not just traditional software engineering teams. While developers remain the primary users, legal and marketing departments have also begun integrating AI tools into daily operations to streamline tasks and improve productivity.
Despite the rapid adoption, Uber executives emphasized that the relationship between AI usage and measurable product improvements is still being evaluated. The company is carefully assessing whether increased reliance on coding assistants is directly translating into better consumer-facing features.
Budget Pressures Force Caps
Alongside its growing reliance on AI tools, Uber has also introduced strict cost controls. The company recently capped employee spending on AI coding tools at around $1,500 per month per tool after exhausting its allocated AI budget earlier in the year.
The limit applies to popular developer platforms such as Cursor and Anthropic’s Claude Code, both widely used for generating and assisting with software development tasks. Employees are able to monitor their usage through internal dashboards and can request special approval if they need to exceed the set cap.
This move highlights the rapidly rising cost of scaling AI across large enterprises. While these tools can boost productivity, they also introduce unpredictable and sometimes significant operational expenses, especially for heavy users.
Mixed Impact on Productivity
Uber leadership noted that AI agents are now responsible for approximately one-tenth of all committed code within the company. However, executives remain cautious about overstating the immediate benefits of this shift.
President and COO Andrew Macdonald pointed out that while usage of tools like Claude Code is increasing, it is not yet clear how directly this translates into improved consumer product performance. In other words, more AI-generated code does not automatically mean better or faster product innovation.
CEO Dara Khosrowshahi has also highlighted the expanding role of AI across different departments, signaling that the company sees long-term potential beyond just engineering. Still, Uber appears to be balancing enthusiasm for AI with a measured approach to its real-world business impact.
Industry-Wide Cost Realignment
Uber’s decision also reflects a broader trend across the technology sector, where companies are beginning to rein in AI-related spending after rapid early adoption. As AI coding tools become more powerful and widely used, their pricing structures have created new challenges for enterprise budgets.
Some platforms, such as Cursor, offer custom enterprise pricing models, but third-party analyses suggest that heavy usage can still lead to significant daily overage charges beyond standard licensing fees. This unpredictability is pushing companies to introduce tighter controls.
Other tech giants are making similar adjustments. Microsoft, for example, has reportedly reduced direct use of Claude Code among engineers and is shifting teams toward GitHub Copilot’s command-line tools as part of its own cost optimization strategy.
Uber’s latest move highlights a critical turning point: AI is no longer just an experimental productivity booster, but a cost-sensitive infrastructure layer that companies must actively manage.


