Uber Burned Through Its AI Budget in Four Months. What Did It Get?

Uber spent $3.4 billion on research and development in 2025 — 9 percent more than the year before, according to The Verge. By April 2026, just four months into the year, the company had already exhausted its annual AI budget.

So what did Uber get for that? According to its own president, the honest answer is: unclear.

"That link is not there yet," Uber president and COO Andrew Macdonald said in an interview with Rapid Response. He was talking specifically about the connection between skyrocketing token consumption for Claude Code and the number of useful features actually shipped to consumers.

"It's very hard to draw a line between one of those stats and, 'Okay, now we're actually producing 25 percent more useful consumer features,'" Macdonald said.

One of the largest tech companies in the world is spending billions — and its own executive can't explain what it bought.

The Headcount Trade-Off Is Real

Uber CEO Dara Khosrowshahi said this month the company is compensating for rising AI costs by hiring fewer human employees, according to The Verge. Translation: people are losing jobs to AI spending, and that AI spending isn't yet provably delivering better products.

Macdonald put it plainly: "We're going to have to start talking about token consumption and the associated cost versus headcount. So if you're not actually able to draw a direct line to how much useful features and functionality you're shipping to your users, that trade becomes harder to justify."

Goldman Sachs and a16z have pushed back on the AI job apocalypse narrative, pointing to historical precedent where technology created more jobs than it destroyed. That argument has merit. But it assumes the productivity gains are real and measurable. Uber is now raising a direct challenge to that assumption.

The productivity dividend from AI isn't guaranteed. It may still come. But right now, at Uber, it isn't showing up in the numbers.

The Mainstream Media Is Missing the Bigger Picture

Left-leaning outlets like The Verge are covering the Uber story as a cautionary tale about AI hype. But they're not asking the harder follow-up question: why is no one being held accountable for burning billions with zero measurable return?

If a government agency blew through its annual budget in four months with nothing to show for it, outlets across the political spectrum would demand answers. When a private tech company does it, it's treated as a learning experience.

Right-leaning media largely isn't covering this angle at all — too busy defending AI as the solution to everything. Both sides are leaving the story incomplete.

The Suno Problem: AI Isn't Just Replacing Workers — It's Replacing Culture

The Uber story is about corporate ROI. There's a parallel story unfolding that's harder to quantify.

According to The Verge, a visible trend has emerged in the Suno subreddit: users aren't just making AI music for fun. They're abandoning streaming platforms entirely and listening almost exclusively to their own AI-generated output.

Quotes from Reddit users cited by The Verge:

• "I definitely listen to my own music most of the time now. Why wouldn't I? It's album after album of bangers."
• "Last.fm says I listened to my own (AI music) 2239 times in the last 365 days."
• "I hardly ever use Spotify bc I favor my own stuff more."

The Verge's Terrence O'Brien reached out to over a dozen of these users to ask why. Nobody would talk on the record. Not one person.

These users know, on some level, that their preference is difficult to defend in conversation with another human being. But they keep doing it anyway.

This Isn't About Music Snobbery

The easy dismissal is to call this an aesthetic debate — maybe AI-generated music just isn't as good as real music, so who cares if a few Reddit users prefer their own output.

The real issue is about displacement. Real musicians — who spent years developing actual craft — are losing listeners not because better music won, but because personalized mediocrity is frictionless.

One user claimed AI gives them "far out genres" like "country/rap" — a genre that Lil Nas X popularized years ago. The issue isn't that human artists haven't made what these users want. The issue is that AI makes it easier to consume something tailored to you than to discover what already exists.

Personalization as a replacement for discovery. That's a cultural problem, not just a tech story.

What This Actually Means

Two separate stories. One theme: AI is replacing things without clearly improving them.

At Uber, it's replacing headcount without demonstrably improving output. In the Suno subreddit, it's replacing real artists without producing objectively better music.

In both cases, the replacement happens because AI is cheaper, faster, or more convenient — not because it's better.

When the trade-off is invisible — when companies quietly swap employees for tokens they can't justify, or when artists quietly lose listeners to algorithms nobody will defend publicly — the stakes change.

The AI productivity story isn't dead. But Uber's own president is asking for evidence. That's the question worth examining.