Uber, Others Say AI Spend Hard To Justify As Token Use Rivals Labor Costs

The whispers are growing louder. While it may still be a career risk to question the benefits of AI, the harsh reality is that more executives at more companies are doing so. It is important to note that they are not challenging AI's potential impact - emphasis on potential: what they are raising is the cost benefit equation. They know the cost - boy, do they - but the benefit? In too many cases, eh, not so much.

And given that one of AI's ostensible features is its ability to supplant human employees, that corporate execs are now stating that the cost of tokens had driven AI costs up to the level of that much maligned human capital, thereby threatening a core premise for AI adoption. Silicon Valley hypesters can flame doubters all they want, but when too many leaders' compensation became threatened, there was always going to be pushback. And that has now begun. JL

Cris Tolomia reports in Quartz and Reed Albergotti reports in Semafor:

Uber's COO said the company is struggling to connect its rising AI costs to meaningful gains in consumer-facing products. Uber's engineering leaders concluded that greater token consumption was not yielding a corresponding rise in features delivered. "It's very hard to draw a line between those and, 'now we're producing 25% more features." In April the company exhausted its entire Claude Code budget for 2026 — just four months into the year. That was a "head-exploding moment" about the cost of tokens and what the company gives up to sustain it. Uber is not alone in reassessing its approach to AI usage. The cost of tokens is now competing with the cost of headcount. "Unit costs are going down but aggregate costs are going up." Free AI models are “the gateway drug” to paid versions. “What should we be spending, and how do we measure ROI?” 

Uber $UBER 0.00%'s chief operating officer said the company is struggling to connect its rising artificial intelligence costs to meaningful gains in consumer-facing products, according to The Verge.  

Andrew Macdonald, Uber's president and COO, made the comments in an interview with Rapid Response released over the weekend. Drawing on conversations with Uber's senior engineering leaders, Macdonald concluded that greater token consumption was not yielding a corresponding rise in useful features delivered to end users. "That link is not there yet, right? I think maybe implicitly there is more that is getting shipped, but 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 told the outlet. 

The remarks come after Uber CTO Praveen Neppalli Naga disclosed in April that the company had exhausted its entire Claude Code budget for 2026 — just four months into the year, according to Business Insider. Macdonald called the CTO's disclosure a "head-exploding moment," saying it set off company-wide conversations about the cost of token consumption and what the company gives up — including potential headcount — to sustain it.

"We're going to have to start talking about token consumption and the associated cost versus headcount," Macdonald told The Verge. "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."

Engineers who are not responsible for paying invoices may treat AI tools as essentially free, Macdonald observed, even though those costs land on the company's balance sheet, according to Business Insider.

Uber disclosed in its first-quarter earnings report that AI coding tools had reached 95% monthly adoption among its engineering workforce, with AI agents authoring more than one in ten lines of code. The company spent $3.4 billion on research and development in 2025, up 9% from the prior year. CEO Dara Khosrowshahi said during the earnings call that Uber was slowing hiring to offset its growing AI investments.

According to Forbes, the share of Uber engineers using Claude Code jumped from roughly a third in February to 84% in March, and per-engineer bills typically ran between $150 and $250 a month, though the heaviest users could generate anywhere from $500 to $2,000. During one hands-on demonstration he ran himself, Neppalli Naga burned through $1,200 worth of tokens in just two hours.

Uber is not alone in reassessing its approach to AI usage. At Duolingo, management reversed course on a policy that had tied employee performance evaluations to AI usage, after staff raised concerns that the metric rewarded tool adoption rather than actual results. "It felt like, rather than being held accountable for the actual outcome, we were trying to just push something that in some cases did not fit," Duolingo CEO Luis von Ahn said in a podcast interview in April, according to Business Insider.  

The cost of tokens is now competing with the cost of headcount.

“You have to raise a lot more money on a per-headcount basis,” Henry Ward, who runs Silicon Valley financial-software provider Carta, said at Semafor World Economy in Washington, DC. “I don’t see any end to that in sight.”

Tech executives like Nvidia’s Jensen Huang have suggested that the more companies spend on AI, the more money they will make. But the growing cost of tokens is a line item that CFOs are having trouble planning for. Several executives in Washington, DC, last week said they’re grappling with that uncertainty.

“The unit costs are going down, but the aggregate costs are going up, and companies don’t like when something is unpredictable on cost,” said Charles Phillips, co-founder of private equity firm Recognize. Kunal Kapoor, CEO of Morningstar, added that subscription models “have endured for a long time because there’s value in certainty.”

ICONIQ Capital founding partner Divesh Makan, whose investment firm recently led one of Anthropic’s funding rounds, called free AI models “the gateway drug” to paid versions, adding that his employees keep asking for more tokens.

“The conversation we’re having is, ‘What is the amount we should be spending, and how do we measure ROI?’” he said. “Are you just buying holidays and checking the weather in Tokyo, or are you doing something productive with these tokens?”