Uber tech chief: The AI tokenmaxxing era is ending

The mood around corporate AI spending is shifting, and Uber is making that sentiment explicit. Praveen Neppalli Naga, the company’s chief technology officer, declares that the era of tokenmaxxing is drawing to a close. For the uninitiated, tokenmaxxing refers to the relentless escalation of AI investment, quantified by the sheer volume of tokens models process. That approach, he argues, is no longer sustainable.

Speaking with clarity that many in the tech sector are only beginning to adopt, Neppalli Naga frames this as a natural maturation. The initial gold rush, where throwing more compute and more data at problems was the default strategy, has given way to a more disciplined mindset. Companies are now prioritizing efficiency and tangible returns over raw scale. The question is no longer how many tokens you can burn, but what value those tokens actually generate.

This pivot is significant coming from Uber, a company that processes massive amounts of data daily across its global ride-hailing and delivery networks. The CTO’s remarks signal a broader recalibration across the industry. Investors have grown wary of open-ended AI budgets, and executives are feeling the pressure to demonstrate a clear link between spending and business outcomes.

Neppalli Naga’s point is not that AI itself is falling short. Rather, the focus is shifting toward smarter application. The days of measuring success by model size or inference volume are fading. What matters now is how well AI integrates into workflows, reduces costs, or opens new revenue streams. That demands a different kind of engineering discipline, one that optimizes for precision rather than volume.

The term tokenmaxxing may be new, but the underlying behavior is familiar. It mirrors earlier tech cycles where companies overbuilt capacity in the hope that usage would catch up. Eventually, reality sets in, and the emphasis moves to operational excellence. Uber appears to be positioning itself ahead of that curve.

For the broader market, the takeaway is clear: AI spending is entering a new phase, one defined by accountability. The companies that thrive will be those that treat AI as a tool for solving specific problems, not as an end in itself. The era of indiscriminate expansion is over, and a more measured, outcome-driven approach is taking its place.