Uber’s chief technology officer, Praveen Neppalli Naga, says the company is moving away from what he describes as a “tokenmaxxing” approach to AI. The term refers to a practice of spending increasingly large amounts on AI systems, often measured in the number of tokens models consume. Naga characterizes this as an era that is coming to an end, suggesting a shift in how Uber evaluates and uses AI resources.
In discussing the company’s AI strategy, Naga highlights that Uber’s adoption of AI has grown rapidly. He says AI adoption at the company has quadrupled, indicating expanding use of AI across teams or products. At the same time, he says Uber is now prioritizing more efficient use of AI rather than simply increasing consumption. This implies a focus on productivity, cost controls, and better performance per unit of AI usage.
Both sources present the same core message: Uber is signaling a change from scaling AI spending toward optimizing AI usage as corporate AI investment priorities evolve.
