Uber president says AI spending is getting ‘harder to justify’
Uber says rising AI costs are getting harder to tie to visible product gains. Its president says the company still sees growth in usage, but not a clear return.
Intelligence analysis by GPT-5.4 Mini
Uber is questioning the payoff from its AI spending after burning through its annual budget early in 2026. The company says token usage is rising fast, but it cannot yet connect that spend to more useful features for riders.
Uber is spending a lot of money on AI, but it is not sure the spending is helping enough. It is like buying lots of ingredients for a kitchen and still not knowing if the meals taste better.
The company can see that the AI tools are being used more and more. But it cannot clearly show that this extra use is turning into better apps or more useful features for riders and drivers.
That is why Uber is being careful. It wants to know if the money it spends is building something people can actually feel, not just making the computer work harder.
Analysis
What Uber is saying
Uber president and COO Andrew Macdonald says the company is struggling to connect its AI spending with measurable product gains. He said it is “hard to draw a line” between rising token usage for Claude Code and the delivery of more useful consumer features.
The core problem
The company appears to be seeing strong underlying AI activity, but not a clear business outcome that justifies the cost. Macdonald said the link between usage stats and something like “25 percent more useful consumer features” is not there yet, even if the raw metrics are moving sharply upward.
Uber’s spending backdrop matters here. The company spent $3.4 billion on research and development in 2025, which the article says was 9 percent more than the year before. Earlier in May, CEO Dara Khosrowshahi said Uber was offsetting higher AI investment by hiring fewer human employees.
Why this is notable
The article frames a familiar tension in AI adoption: usage can rise quickly while returns stay fuzzy. Uber is not rejecting AI spending outright; instead, it is saying the trade-off is getting harder to defend unless the company can show how token consumption translates into actual user-facing value.
That makes this story less about one company’s budget and more about a broader test facing AI-heavy firms: whether the bills being paid for models and tokens will reliably produce products customers can notice.
Key points
- Uber says AI spending is becoming harder to justify because the company cannot clearly link it to better features.
- Andrew Macdonald said rising token usage does not yet translate into measurable consumer value.
- Uber spent $3.4 billion on R&D in 2025, which was 9 percent more than the prior year.
- CEO Dara Khosrowshahi recently said Uber is balancing AI investment by hiring fewer people.
- The story highlights a broader question for AI companies: whether usage growth is producing real business returns.
