Nvidia Scaled Robotaxi Technology Across Global Fleets
The company has expanded its autonomous computing stack, which now supports major commercial robotaxi operators.
Updated on Sept. 25, 2026 in Robotics
Live Poll
Do you trust self-driving technology enough to use it for your daily transportation needs?
Nvidia has announced its continued role as the primary technology provider for global robotaxi programs, utilizing a stack that spans AI training, simulation, and in-vehicle processing. The platform aims to solve for rare driving scenarios that are difficult to capture in real-world data collection.
Why it matters
By providing standardized hardware and simulation tools, Nvidia aims to accelerate the deployment of autonomous commercial vehicles, a market projected to reach 6 million units by 2035. Its approach addresses the persistent bottleneck of capturing dangerous edge cases in training data.
The DRIVE Hyperion 10 architecture integrates dual DRIVE AGX Thor chips with a sensor suite comprising 14 cameras, nine radars, three lidars, and 12 ultrasonic sensors. This configuration demonstrated a 43% improvement in trajectory prediction error during internal testing.
The players
Nvidia
A designer of graphics processing units and AI-optimized data center hardware that provides the computing foundation for autonomous vehicle training and inference.
Uber
A mobility-as-a-service company developing an AI data factory for autonomous vehicle workloads using Nvidia technology.
May Mobility
An autonomous vehicle operator currently running robotaxi services on the Lyft network in Atlanta.
Tesla
An automaker that utilizes Nvidia supercomputers to train neural networks for its proprietary autonomous driving systems.
The details
Nvidia employs a three-part compute strategy consisting of data center-based training, virtual simulation, and real-time in-vehicle processing. The system uses a 'data factory' approach to generate synthetic training examples that mirror rare or hazardous driving conditions. The Hyperion 10 architecture facilitates these tasks by coordinating massive sensory inputs—lidar, which uses laser light to map surroundings, and radar, which measures distances using radio waves—to feed the autonomous driving neural networks.
Timeline
2035: Projected year for the robotaxi market to reach USD $400 billion.
The Tech Race
Nvidia is positioning its DRIVE Hyperion 10 architecture to serve as the dominant infrastructure layer for autonomous transit, directly challenging bespoke internal solutions at other automakers. This represents a critical pivot toward capturing the entire robotaxi stack as the industry moves from research prototypes to mass commercial scale.
Users in markets like Atlanta may see increased availability of robotaxi services as operators like May Mobility scale their fleets on the Lyft network. The technology enables these vehicles to better handle rare, high-risk driving scenarios, potentially improving safety reliability over time.
The takeaway
Nvidia is successfully shifting its business model to become the essential utility provider for the autonomous transit economy. Observers should monitor whether this hardware standardization yields the projected USD $400 billion market size by 2035 as global commercial deployments accelerate.
Further reading
For more on the hardware driving autonomous fleets, visit our Robotics section.
Source note: This article includes information reported by IT Brief Australia.
Live Poll
Do you trust self-driving technology enough to use it for your daily transportation needs?
