Uber burned its 2026 AI budget in 4 months as COO questions ROI

Uber COO Andrew Macdonald says the link between rising AI token spending and useful consumer features "is not there yet."

Uber exhausted its entire 2026 budget for AI coding tools by April after incentivizing employees through an internal usage leaderboard, prompting its chief operating officer to question whether the spending is translating into consumer value.

"That link is not there yet," Andrew Macdonald, president and chief operating officer at Uber, said on the Rapid Response podcast. "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.'"

Uber spent $951 million on research and development in the first quarter of 2026 alone, a nearly 17 percent increase from the same period a year earlier. The company's R&D spending represented 3.4 percent of revenue in 2025, up 9 percent from 2024. Chief executive Dara Khosrowshahi said on an earnings call that about 10 percent of the company's committed code is now built by autonomous agents.

The tension at Uber reflects a broader reckoning across corporate America. Gartner forecasts AI agent software spending will reach $207 billion in 2026, up 139 percent from $86.4 billion in 2025. But a growing number of executives are struggling to connect those costs to measurable returns. "If a year from now we're still talking about the same question, I'd be more concerned," OpenAI chief executive Sam Altman said recently.

A Leaderboard-Driven Culture

Uber's approach to AI adoption included an internal leaderboard ranking teams by total usage of tools such as Anthropic's Claude Code. The strategy worked — perhaps too well. By April, the company had already exhausted its Claude Code budget for the full year, according to chief technology officer Praveen Neppalli Naga, as reported by The Information.

The phenomenon has drawn criticism from investors and analysts. Michael Burry, who famously bet against the 2008 housing bubble, called the trend "quota-driven, leaderboard-driven, management-mandated overconsumption" in a Substack post. Shruti Gandhi, a general partner at Array Ventures, compared tokenmaxxing to a factory CEO bragging about the power bill while all the machines idle. "Spending more doesn't mean producing more," she wrote on X.

The Industry's ROI Reckoning

Uber is far from alone. Microsoft earlier this month began canceling most of its direct Claude Code licenses, moving engineers toward GitHub Copilot CLI, according to The Verge. Duolingo chief executive Luis von Ahn last year reversed his earlier AI-first stance, saying he no longer sees the technology replacing the tasks his employees perform.

Google chief executive Sundar Pichai acknowledged hearing directly from chief information officers who are "so concerned about how much their companies are blowing through budgets." His response was to pitch a more efficient model — Gemini 3.5 Flash — rather than dismiss the concerns.

Even OpenAI's Altman, whose company benefits directly from rising token consumption, said the question of ROI is top of mind for corporate leaders. "There's a lot of great things I hear from companies, the negative one I hear is 'our spending is going up and up, people feel like they're being very productive… but where is the revenue, where are the actual productivity gains?'" Altman said during a virtual appearance at an AI event in Australia.

For investors, the growing skepticism around AI ROI introduces a new variable into the valuation of companies across the AI supply chain. Uber shares have yet to reflect the spending debate — the company trades at roughly 22 times forward earnings, with analysts focused more on its autonomous driving ambitions than its AI coding costs. But if the pattern spreads, companies that have built their growth narratives around AI adoption — from cloud providers to chipmakers — may face tougher questions about whether their customers' spending is sustainable. Macdonald himself said Uber is going all in on autonomous driving, a technology he expects will make driver's licenses obsolete for his children. The question is whether the AI spending required to get there will face increasing scrutiny along the way.

This article is for informational purposes only and does not constitute investment advice.