Launch begins in Los Angeles and San Francisco in 2027
Uber Technologies and NVIDIA announced an expanded autonomous vehicle partnership aimed at deploying a global fleet of autonomous vehicles powered entirely by NVIDIA software. The companies said service is expected to begin in Los Angeles and San Francisco in the first half of 2027, with plans to scale to 28 cities globally by 2028.
NVIDIA platform and new AI model sit at the center
The rollout will be built around the NVIDIA DRIVE Hyperion autonomous vehicle platform and NVIDIA Alpamayo, a next-generation reasoning-based AI model intended to handle complex driving edge cases such as unpredictable construction zones and erratic pedestrian behavior. NVIDIA positioned Alpamayo as part of its move toward becoming a full-stack Level 4 software provider.
Phased deployment strategy for each city
Uber and NVIDIA said each launch market will follow a staged approach:
- Data-collection fleet to train Alpamayo on local driving patterns and nuances
- Operator-led launch as service expands in-market
- Fully driverless Level 4 deployment as the final phase
The companies said this structure is designed to support scaling driverless operations across North America, Europe, Australia, and Asia by 2028.
How Uber frames commercialization
Uber CEO Dara Khosrowshahi said the partnership combines NVIDIA’s AI capabilities with Uber’s network and operating experience, with the goal of broad commercialization and a multi-operator autonomous ecosystem over time.
NVIDIA positions robotaxis as a physical AI inflection point
NVIDIA CEO Jensen Huang described robotaxis as part of a broader shift toward “physical AI,” arguing that robotic systems can increasingly reason about real-world complexity. He also emphasized connecting NVIDIA’s ecosystem of robotaxi-ready partners to Uber’s ride-hailing platform.