Uber and NVIDIA Plot a Real Road for Robotaxis

Uber and NVIDIA team up to bring large autonomous fleets from test tracks to U.S. streets

9 Dec 2025

The effort to bring autonomous vehicles into everyday transportation is entering a new phase, as Uber and Nvidia align their platforms to support large-scale deployment in U.S. cities in the coming years. The partnership is aimed less at experimentation and more at answering a question that has long challenged the industry: how to move from limited trials to sustainable operations.

For much of the past decade, autonomous driving has been defined by demonstrations in carefully managed environments. Those pilots helped validate the technology, but they did little to resolve the operational, regulatory and economic hurdles of running driverless vehicles day after day. The Uber-Nvidia collaboration reflects a broader recognition that proving autonomy works is no longer enough. Scaling it reliably and safely has become the central task.

Uber’s role draws on its experience running a vast transportation network. The company manages millions of trips daily and has built systems for fleet coordination, pricing and compliance across cities with differing rules. Rather than developing its own autonomous driving technology, Uber has positioned itself as a platform where vehicles developed by partners can be deployed, managed and eventually monetized at scale, according to company statements.

Nvidia provides the technical foundation. The partnership centers on Nvidia’s DRIVE AGX Hyperion platform, which combines in-vehicle computing hardware with software designed for advanced automated driving. Over time, the companies say, the system could support as many as 100,000 Level 4 autonomous vehicles. Nvidia has argued that access to large volumes of real-world driving data is essential to improving safety and performance, a point its chief executive, Jensen Huang, has emphasized in public remarks.

Vehicle suppliers are also part of the strategy. Stellantis, among others, plans to offer models engineered to support autonomous systems, using standardized, autonomy-ready designs. Such an approach could reduce integration costs and ease expansion, problems that have slowed previous attempts at commercialization.

The implications for riders and cities remain uncertain and incremental. Widespread deployment could eventually lower ride prices and improve availability, analysts said, but regulatory scrutiny, public safety expectations and operating costs continue to pose challenges.

Still, the partnership signals a shift in how the industry views progress. Autonomous mobility’s next chapter appears less about isolated technical breakthroughs and more about coordinated efforts that link technology, vehicles and operations. That approach could shape transportation policy and markets in the years ahead.