NVIDIA DRIVE Hyperion Lands BYD, Geely, Nissan as L4 Autonomy Race Heats Up

Felix Pinkston Mar 16, 2026 22:24

NVIDIA announces major automaker partnerships for Level 4 autonomous vehicles, with Uber robotaxi deployment across 28 cities by 2028 starting in LA and SF.

NVIDIA DRIVE Hyperion Lands BYD, Geely, Nissan as L4 Autonomy Race Heats Up

NVIDIA just locked in four major automakers for its Level 4 autonomous driving platform, signaling the company's aggressive push to become the backbone of the robotaxi industry. BYD, Geely, Isuzu, and Nissan will build L4-ready vehicles on NVIDIA DRIVE Hyperion, the company announced at GTC on March 16.

The bigger news? NVIDIA and Uber are launching full-stack robotaxis across 28 cities on four continents by 2028. Los Angeles and San Francisco get first dibs in the first half of 2027.

What DRIVE Hyperion Actually Delivers

The platform bundles compute, sensors, networking, and safety systems into a standardized architecture. For automakers, that means faster validation cycles and simpler global deployment. The system runs on DRIVE AGX Thor—built on NVIDIA's Blackwell architecture—and integrates with the new Halos OS safety framework.

Jensen Huang didn't mince words during the keynote: "The autonomous vehicle revolution is here—the first multitrillion-dollar robotics industry. Everything that moves will eventually be autonomous."

That's bold, but the partnership roster backs it up. Beyond the headline automakers, mobility players Bolt, Grab, Lyft, and TIER IV are also building on Hyperion. Isuzu and TIER IV specifically target L4 autonomous buses using the platform.

The Software Stack Gets Smarter

NVIDIA dropped Alpamayo 1.5 alongside the hardware news. The open model takes driving video, motion history, navigation data, and natural language prompts as inputs, then outputs trajectories with reasoning traces. Developers can steer behavior through text prompts—useful for handling edge cases like unusual road hazards.

Over 100,000 automotive developers have downloaded Alpamayo since launch earlier this year. The 1.5 upgrade adds multi-camera support and configurable parameters, letting teams reuse the same AI stack across different vehicle lines.

For simulation, NVIDIA Omniverse NuRec uses 3D Gaussian Splatting to reconstruct real-world environments for testing. Partners including dSPACE, Foretellix, and the University of Michigan's Mcity facility have already integrated it.

Market Position and What's Next

NVIDIA shares traded at $180.20 on March 16, down 1.58% on the day despite the announcement—likely reflecting broader market conditions rather than the news itself. The company's market cap sits at $4.5 trillion.

The Uber timeline provides concrete milestones to watch. LA and SF deployments in H1 2027 will test whether NVIDIA's full-stack approach can compete with Waymo's established California presence. The 28-city target by 2028 represents aggressive scaling that could reshape urban transportation economics.

Hyundai Motor and Kia also expanded their NVIDIA partnership the same day, adding to the automaker coalition. With safety certifications aligned to ISO 21434 and ISO 26262 standards, NVIDIA is positioning Hyperion as the path of least resistance for OEMs racing toward autonomy.

Image source: Shutterstock