• Early surge in AI token spending
• Macdonald’s “hard to justify” moment
• Wider Silicon Valley backlash and bubble fears
• Uncertain path forward

Uber Executive Questions Return on Investment of AI Spending Uber’s aggressive push into artificial intelligence is colliding with financial reality, as one of Silicon Valley’s biggest spenders now admits it can’t clearly see what all that compute is buying.

§Early surge in AI token spending

Over the past year, major tech firms have raced to “tokenmaxx” — pushing ever more AI usage, often tracking staff by how heavily they lean on tools like chatbots and code assistants. Visa, Disney, JPMorgan and others have publicly highlighted massive token consumption as a sign of innovation, while internal AI usage became a performance signal at some firms.

Uber joined that wave, pouring money into AI research and development — $3.4 billion in 2025, up 9 percent year over year. By early 2026, the company had reportedly exhausted its entire annual AI budget just four months into the year, largely on Claude Code usage.

§Macdonald’s “hard to justify” moment

The rethink crystallized after Uber CTO Praveen Neppalli Naga disclosed that the company had already blown through its Claude Code budget for 2026, a revelation that Uber’s operations chief Andrew Macdonald described as a “head-exploding moment.” In a subsequent interview, Macdonald said rising token consumption was not translating into proportionally more “useful consumer features.” “That link is not there yet,” he said, adding that it is “very hard to draw a line” between token stats and, for example, “25% more useful consumer features.”

Uber CEO Dara Khosrowshahi had already warned that AI investments were forcing slower hiring, explicitly trading human headcount for AI spend. Macdonald echoed that pressure: “We’re going to have to start talking about token consumption and the associated cost versus headcount,” he said, noting that without a clear link to shipped features, “that trade becomes harder to justify.”

§Wider Silicon Valley backlash and bubble fears

Macdonald’s comments landed in an ecosystem already uneasy about costs. Business Insider described a “backlash against tokenmaxxing” as companies realize they may be “spending billions on AI tokens and nobody can agree if it’s working.” Google CEO Sundar Pichai has said CIOs are “so concerned about how much their companies are blowing through budgets” on AI and warned the problem is likely to worsen.

Some practitioners argue that a significant share of internal token spend is wasteful, with engineering leaders complaining that “tokens got burned for millions of dollars without any real significant ROI to show for it.” Others, like Duolingo, have begun rolling back policies that reward AI use for its own sake, after employees questioned whether they were being pushed to deploy AI even when it “did not fit.”

§Uncertain path forward

For now, Macdonald suggests the benefits of heavy AI use may emerge only over time, saying that “over the coming quarters and years, maybe that will become clearer,” even as underlying AI metrics look “astronomical.” The tension for Uber — and for Silicon Valley at large — is how long they can justify that bet without hard evidence that more tokens truly mean better products.

Continue reading https://foxvector.com