DooDooLamb News
Uber president says AI spending is getting ‘harder to justify’
Brief published May 28, 2026 · Original source published May 26, 2026
Original reporting by Jess Weatherbed at theverge.com.
Automated brief. Verify important details at the original source.
Uber's AI Budget Goes Missing, Company Shocked to Discover Money Doesn't Automatically Equal Results
Uber, the company that spent over a decade perfecting the art of burning investor cash while calling it "growth," has apparently run through its entire 2026 AI budget in just four months and is now experiencing what economists call "buyer's remorse" but what the rest of us call "finally reading the receipt."
President and COO Andrew MacDonald recently admitted in an interview that the ride-hailing giant is struggling to find a "clear connection between AI usage and productivity," a revelation roughly equivalent to Columbus discovering that the Earth might not actually be flat. The confession comes as Uber joins a growing chorus of companies realizing that throwing money at AI might not be the same thing as having a strategy for it.
The timing couldn't be more perfect. Here we are, just four months into 2026, and Uber has already spent their entire annual AI allocation. This is impressive even by Silicon Valley standards, where burning through budgets is considered a core competency. It's like eating your Halloween candy before Labor Day, except instead of a sugar crash, you get a boardroom full of executives asking uncomfortable questions about ROI.
MacDonald's admission that there's no clear productivity boost from AI spending reads like the tech industry's equivalent of "we tried everything and we're all out of ideas." This is particularly rich coming from a company whose entire business model was initially "what if taxis, but with an app?" Now they're discovering that "what if everything, but with AI?" might not be the innovation goldmine they hoped for.
The broader context here is delicious. Uber has spent years positioning itself as a technology company rather than a taxi dispatcher with venture capital backing. They've weathered scandals, regulatory battles, and the minor detail of never actually turning a consistent profit until relatively recently. Now, having finally figured out how to make money moving people and food around, they've apparently decided to stress-test that profitability by seeing how quickly they can spend it on the hottest trend in tech.
This AI budget exhaustion isn't happening in a vacuum. Companies across the tech landscape are discovering that implementing AI is less like installing a software update and more like performing surgery while the patient is running a marathon. The promise was simple: sprinkle some machine learning on your operations and watch efficiency soar. The reality has been more like hiring a very expensive intern who occasionally produces brilliant work but mostly just asks a lot of questions and requires constant supervision.
What makes Uber's situation particularly amusing is that their core business already relies heavily on algorithmic optimization. Their entire platform is built around matching drivers with riders, optimizing routes, and predicting demand patterns. These are fundamentally AI-adjacent problems that they've been solving for years. But apparently, having working AI solutions for your actual business problems doesn't count unless you can slap "generative AI" or "large language model" on the marketing materials.
The company's AI spending spree likely included all the greatest hits: chatbots for customer service that understand every language except the one angry customers actually speak, predictive analytics that can forecast everything except whether a driver will actually show up, and probably some sort of AI-powered route optimization that works perfectly until it encounters the exotic edge case of "traffic exists." Each of these initiatives probably came with a PowerPoint deck promising revolutionary improvements and a price tag that would make a defense contractor blush.
What's particularly telling is MacDonald's struggle to quantify productivity gains. This suggests that Uber, like many companies, fell into the classic trap of implementing AI solutions in search of problems to solve, rather than the other way around. It's the corporate equivalent of buying a really expensive hammer and then wandering around the office looking for things that might need hammering.
The reality check here is both sobering and oddly comforting for anyone who's been watching the AI hype cycle with skepticism. Uber's experience suggests that even companies with legitimate technical expertise and clear use cases for algorithmic solutions can struggle to extract value from AI investments when they're driven more by FOMO than by strategic thinking.
This doesn't mean AI is useless, any more than the dot-com crash meant the internet was a fad. But it does mean we're hitting the inevitable "trough of disillusionment" where companies start asking hard questions about whether their AI initiatives are actually solving real problems or just burning money in increasingly sophisticated ways. The fact that Uber, a company built on algorithmic optimization, is having this moment should give pause to every startup currently pivoting to "AI-first" whatever-they-were-doing-before.
For developers and founders watching this unfold, Uber's confession is a gift. It's permission to stop pretending that every product needs an AI feature and every budget needs an AI allocation. It's a reminder that the most revolutionary thing you can do in 2026 might just be building something that works well, rather than something that works well with artificial intelligence.
The funniest part of all this is that somewhere in Uber's offices, there's probably a spreadsheet tracking their AI ROI that was generated by AI, analyzing data collected by AI systems, and nobody can quite explain why the numbers don't add up. Sometimes the future arrives exactly as promised, just not in the way anyone expected.