Nvidia Partners with BYD and Geely for Robotaxi Technology in China

Nvidia has welcomed two prominent Chinese automakers, BYD and Geely, into its robotaxi initiative, aiming to make a significant impact in the expanding global autonomous vehicle sector. During its GTC conference, Nvidia revealed that BYD, Geely, along with Isuzu and Nissan, will utilize the Drive Hyperion platform, which integrates the necessary chips, computers, sensors, and software for developing Level 4 autonomous vehicles.

BYD currently employs Nvidia’s chips in its conventional vehicles, and this new partnership will enable it to leverage the Hyperion platform for its next-gen Level 4 models. Geely is reportedly using Nvidia’s Thor chips in its latest Zeekr vehicles, which are also supplied to Waymo for its robotaxi service in the US. According to Nvidia’s VP and general manager of the automotive team, Waymo is utilizing Nvidia’s technology both in vehicles and cloud services.

This collaboration comes at a time of heightened trade tensions between the US and China, particularly concerning technology and tariffs. Nvidia’s chips, especially those used in data centers for AI training, have been central to negotiations between the two nations, with recent approvals allowing the sale of Nvidia’s H200 chips to Chinese firms.

China currently leads the US in electric vehicle production, but both countries are closely matched in the robotaxi arena. For instance, Baidu operates commercial robotaxis in over a dozen cities in China, while Waymo has around 3,000 vehicles in operation across 10 US cities.

Nvidia’s partnerships with BYD and Geely could significantly speed up the development of autonomous vehicles in China, potentially giving the country an edge over the US. Some lawmakers in Congress are advocating for the passage of long-delayed autonomous vehicle legislation to maintain a technological advantage over China.

While Nvidia has long been a supplier of chips and software for driver-assist systems, its automotive sector remains relatively small compared to its AI revenue. In the third quarter of 2025, Nvidia reported total revenues of $51.2 billion, with its automotive division contributing only $592 million, or 1.2 percent of the overall earnings.

Nvidia is not limiting its AV technology to Chinese companies; it will also provide its Hyperion platform to Nissan, which is developing robotaxi software in collaboration with Wayve. Additionally, Nvidia is partnering with Isuzu and Tier IV to create Level 4 buses using its next-gen Drive AGX Thor-based system-on-a-chip.

Moreover, Lyft plans to utilize Nvidia’s Hyperion platform for its own robotaxi development, incorporating vehicles and software from various providers. Lyft also aims to enhance its machine learning capabilities and streamline operations in ridesharing and autonomous vehicle scaling using Nvidia’s technology.

Nvidia has an existing partnership with Uber, with both companies working towards launching a global robotaxi network, targeting a fleet of 100,000 vehicles by 2027. Recently, Nvidia updated this agreement, expanding it to 28 markets across four continents by 2028, with Los Angeles and San Francisco set to be among the first cities in early 2027. Uber is collaborating with several automakers, including Lucid, Volkswagen, and Stellantis, to develop autonomous vehicles powered by Nvidia’s technology.

Kani highlighted Nvidia’s virtual testing capabilities and its open-source AI model portfolio, Alpamayo, as key factors in their progress, despite not having the extensive real-world mileage advantages of competitors like Waymo and Tesla. He noted, “Our AV stack took us more than 10 years to build, but our generational leap was sparked by generalist reasoning models like Alpamayo, as well as synthetic data generation and test capabilities based on Omniverse, NuRec, and Cosmos.”

When asked about the number of companies developing Level 4 vehicles using Nvidia’s products, Kani confidently stated, “I think it’s pretty much everyone,” mentioning a range of companies including Waymo, Nuro, Waabi, Zoox, Wayve, Momenta, Pony, WeRide, Baidu, DeepRoute, and ZYT. “I’m super proud of that,” he added.

As robotaxis become more prevalent, safety concerns are rising. Tesla’s Level 2 vehicles have been involved in numerous accidents, resulting in injuries and fatalities, while Waymo’s vehicles have faced issues with traffic laws and congestion.

To address these safety challenges, Nvidia is introducing Halos OS, designed to assist its Level 4 partners in creating secure systems. Kani described Halos OS as a “safety guardrail” for autonomous driving systems that will intervene if AI models are about to make unsafe decisions. “We help our partners build a safe architecture that we can decompose down for every function and ensure that if any computer or sensor fails, the system is still designed to take you to a safe place,” he explained.