Uber's COO says it's getting harder to justify the company's AI spend: 'That link is not there yet'

Key Points:

• Uber COO Andrew Macdonald expressed skepticism about the direct link between increased use of AI coding tool Claude Code and the production of more useful consumer features, highlighting difficulty in justifying rising AI costs without clear output benefits.
• Uber exhausted its entire 2026 AI coding tools budget within four months after incentivizing AI adoption among employees, reflecting the broader challenge of balancing increased AI usage with escalating expenses despite falling per-unit AI pricing.
• Other tech companies, including Microsoft, are also reevaluating AI tool usage, with Microsoft shifting engineers from Claude Code to GitHub Copilot CLI, while some leaders like Duolingo’s CEO have tempered their initial optimism about AI’s impact on job functions.
• Uber CEO Dara Khosrowshahi noted that about 10% of the company’s code is generated by autonomous agents and emphasized AI’s broader adoption across teams, but rising AI costs remain a concern as industry pricing models evolve towards usage-based billing.
• Despite cost challenges, Uber continues to invest heavily in innovation, particularly autonomous driving, with Macdonald predicting it will become mainstream within two decades and potentially eliminate the need for traditional driver’s licenses for the next generation.