Uber president says AI spending is getting ‘harder to justify’

Uber's president just said the quiet part out loud. After burning through their entire 2026 AI budget by April, the company is struggling to justify what they're getting for all that spending.

Andrew Macdonald, Uber's president and COO, told Rapid Response that the company can't draw a clear line between their AI investments and features that actually matter to users. Specifically, he called out rising token consumption for Claude Code, noting they're spending more but not seeing proportional improvements in what gets shipped to customers.

This is the conversation happening behind closed doors at a lot of companies right now. AI tools promise productivity gains, but measuring actual output is messy. Macdonald acknowledged there might be more getting shipped implicitly, but admitted it's hard to prove they're delivering 25 percent more useful features despite the increased AI spend.

For anyone managing AI budgets or advocating for AI tools at work, this matters. The honeymoon phase is over. Finance teams want proof that AI subscriptions and API costs are translating to real business value, not just busy work that feels productive.

Uber isn't alone in this. As AI costs scale with usage rather than seats, companies are realizing their spending can balloon quickly without clear metrics tying it to outcomes. The pressure is now on to show AI tools are worth the investment, not just nice to have.

The takeaway: if you're using AI tools professionally, start documenting concrete wins. Vague productivity claims won't cut it anymore when budget conversations happen.