Turns out throwing buckets of money at AI doesn't automatically make your app better. Who knew. (Everyone. Everyone knew.)
Uber - the company that charges you surge pricing during a light drizzle - reportedly blew through its entire annual AI budget in just four months into 2026. And now, according to an interview with Rapid Response, Uber president and chief operating officer Andrew Macdonald is starting to ask some very uncomfortable questions about it.
The awkward 'but what did we actually get?' moment
Macdonald told Rapid Response that Uber is struggling to draw a clear line between what it's spending on AI and what features are actually landing in users' hands. In other words: the receipts are enormous, the deliverables are... not.
This is the Silicon Valley equivalent of spending $4,000 on a Peloton in January and then watching it collect dust by March - except the Peloton costs a little less and at least looks good in the corner.
Macdonald reportedly said it's getting "harder to justify" AI spending when there's no obvious connection between rising token costs and real, tangible improvements. Which is a remarkably honest thing to say out loud, in public, as the president of a major tech company during AI's peak hype era.
Why this actually matters
Uber isn't some scrappy startup dazzled by shiny chatbots. It's a global logistics giant with billions in revenue. If they're struggling to see ROI on AI investment, that's a pretty loud signal that the "just spend more and magic will happen" strategy has some fundamental flaws.
The broader tech industry has been on a wild AI spending bender, with companies racing to pour money into infrastructure, models, and integrations - sometimes without a clear plan for how any of it translates into products people actually want. Uber apparently hit that wall at speed, four months into the year.
So... now what?
That's genuinely unclear, and that's kind of the point. The fact that a company as data-obsessed as Uber can't cleanly connect spending to outcomes suggests the whole industry might need a serious reality check - not just a vibe check.
Maybe the real AI disruption was the awkward finance meetings we had along the way.
