Robotaxi and Autonomous Vehicle Solutions
Build safer robotaxis and Level 4 AVs on the industry-standard platform for sensors and compute.
Overview
The next generation of transportation is autonomous, transforming how the world moves, from personal vehicles to robotaxis.
NVIDIA Hyperion™ is a production-ready sensor and compute platform for robotaxis and Level 4-capable autonomous vehicles. It includes standardized sensor suite with dual NVIDIA DRIVE AGX™ Thor system-on-a-chip (SoCs). Combined with NVIDIA Alpamayo Stack, and NVIDIA Halos for safety across all layers of the development lifecycle.
With Hyperion and Halos, teams can focus on building differentiated AI while reducing integration risk, shortening development cycles, and improving safety.
NVIDIA Hyperion brings together the essential capabilities required to build, validate, and deploy next-generation robotaxis and L4-capable autonomous vehicles.
A standardized sensor suite and NVIDIA DRIVE AGX compute are pre-validated to work together, including cameras, radar, lidar, ultrasonics, and in-cabin sensors. Reuse the same architecture across vehicle lines to reduce integration time and cost.
Redundant compute, sensors, and functional safety mechanisms meet all ISO 26262 ASIL-D requirements. When integrated with NVIDIA Halos, NVIDIA Hyperion delivers full-stack safety validation from cloud to car across millions of scenarios.
A single architecture supports passenger vehicles, robotaxis, shuttles, and delivery fleets across global markets. Software-defined and OTA-capable, NVIDIA Hyperion adapts to evolving regulations and use cases throughout a vehicle's lifetime.
Technologies
NVIDIA Halos full-stack safety system and NVIDIA Alpamayo reasoning-based driving software, combined with model development and simulation tools, advance safe AV deployment.
Automotive Partners
Paving the path towards a safer, more intelligent ecosytem.
Resources
Explore the latest announcements, sessions, and videos from the NVIDIA ecosystem.
NVIDIA offers the only robotaxi and AV platform that spans the full end-to-end workflow, from data to deployment, built on safety, modularity, openness, and performance. NVIDIA Halos provides a safety-certified foundation across hardware, software, and sensors, establishing a clear path to regulatory approval. NVIDIA Hyperion is a modular hardware architecture that provides scalability from L2++ driver assistance to Level 4 and flexibility across vehicle programs through configurable compute and sensors. NVIDIA's open models, frameworks, and datasets give every ecosystem partner an adaptable foundation to inspect, extend with proprietary data, and take to production on their own infrastructure. NVIDIA DRIVE AGX Thor and Orin deliver the in-vehicle compute performance to run reasoning models in real time.
NVIDIA Hyperion uses one common architecture across every autonomy level. The compute, sensor suite, and software foundation stay the same from Level 2 through Level 4, and higher levels of autonomy are unlocked through software and over-the-air updates rather than new hardware. The same platform supports driver assistance, robotaxis, shuttles, and delivery fleets, and it gains new capability over the vehicle's life.
NVIDIA Alpamayo is an open family for reasoning-based autonomous driving, including Vision Language Action models, frameworks, and datasets. Alpamayo VLA models process multi-camera video and driving context to generate both a trajectory and a chain-of-causation reasoning trace. Reasoning-based driving improves safety because the vehicle can work through rare situations it was never explicitly trained on, and the reasoning trace makes each decision transparent and auditable for safety validation.
Talk to an NVIDIA automotive specialist about taking your autonomous vehicle program from training to deployment on NVIDIA DRIVE.
Use NVIDIA's open-source models, tools, frameworks, and datasets to take autonomous driving from idea to the road.
Get the latest news on NVIDIA automotive technologies, breakthroughs, and more.