Nvidia added two leading Chinese automakers, BYD and Geely, to its robotaxi program, as the chipmaker seeks to put its stamp on the growing autonomous vehicle market worldwide.
At its GTC conference today, Nvidia announced that BYD and Geely, as well as Isuzu and Nissan, would use the chipmaker’s Drive Hyperion platform, which combines the chips, computers, sensors, and software needed for for the development of Level 4 autonomous vehicles.
BYD currently uses Nvidia’s chips in its manually driven cars, and now, under this expanded agreement, it will use the company’s Hyperion platform to build next-generation Level 4 vehicles. Geely, meanwhile, is said to be using Nvidia’s Thor chips in its new Zeekr vehicles. Geely also supplies Zeekr vehicles to Waymo for its US-based robotaxi service. And Waymo also is using Nvidia’s products, according to Ali Kani, VP and general manager of the automotive team at Nvidia, in a briefing with reporters. “Waymo is using us in the car and the cloud,“ Kani said.
The news that Nvidia is supplying chips and software to two of China’s top automakers comes amid ongoing tensions between the US and China over trade and tariffs. The company’s chips, especially the ones used in data centers to train AI models, have been the subject of intense negotiations between the two countries, with the Trump administration recently approving the sale of Nvidia’s H200 chips to Chinese companies.
China clearly has a huge lead over the US in electric vehicle production, but the two countries are more evenly matched in the robotaxi field. Baidu, for example, is operating commercial robotaxis in over a dozen Chinese cities. (Waymo has approximately 3,000 vehicles operating commercially in 10 US cities.)
Nvidia’s deal with BYD and Geely could vastly accelerate those companies’ development of autonomous vehicles, increasing China’s chances of overtaking the US. Some lawmakers in Congress have pushed to pass long-stalled autonomous vehicle legislation, largely on the premise of maintaining a technological lead over China.
Nvidia is seeking to raise its profile as a self-driving leader. But while the company has long supplied major automakers with chips and software for driver-assist systems, its automotive business is still relatively tiny compared to the billions it rakes in from AI. Its third quarter revenues in 2025 were $51.2 billion, but its automotive division only made $592 million, or 1.2 percent of the total haul.
But Nvidia isn’t just supplying its AV technology to Chinese companies. It also will sell its Hyperion platform to Nissan, which is also using robotaxi software developed by Wayve. And it is working with Isuzu and Tier IV to design Level 4 buses using its next-gen Drive AGX Thor-based system-on-a-chip.
In addition, Nvidia said that Lyft will use its Hyperion platform to develop its own robotaxis, using vehicles and software provided by different companies. Lyft also said it would use Nvidia’s technology to bolster its “machine learning capabilities and accelerate enterprise operations across ridesharing and autonomous vehicle scaling.”
Nvidia already has a partnership with Lyft’s main rival, Uber. The two companies are working together to launch a global network of robotaxis, aiming to deploy a fleet of 100,000 vehicles by 2027. On Monday, the company provided an update on that agreement, saying it now encompasses 28 markets across four continents by 2028 — with Los Angeles and San Francisco to come first in early 2027. Uber is working with multiple automakers, including Lucid, Volkswagen, and Stellantis, who are building autonomous vehicles using Nvidia’s products.
Kani cited Nvidia’s virtual testing capabilities, as well as its open-source portfolio of AI models called Alpamayo, as the reason it’s been able to make so much progress despite lacking much of Waymo’s and Tesla’s advantages in real-world miles driven.
Nvidia is seeking to raise its profile as a self-driving leader.
“Our AV stack took us more than 10 years to build,” Kani said, “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.”
During the briefing, Kani was asked how many companies are developing their Level 4 vehicles on Nvidia’s products. “ I think it’s pretty much everyone,” he said, name-checking Waymo, Nuro, Waabi, Zoox, Wayve, Momenta, Pony, WeRide, Baidu, DeepRoute, and ZYT. “I’m super proud of that,” he added.
With more and more robotaxis hitting the road, people are understandably on edge about safety. Tesla’s Level 2-enabled vehicles have been involved in hundreds of crashes, including 23 injuries and at least two fatalities. Waymo’s vehicles have been recorded violating traffic laws around school buses, and occasionally get stuck in intersections, causing huge traffic jams.
Nvidia’s answer to the problem of these so-called edge cases is to roll out a new product called Halos OS to help its Level 4 partners to build safe systems. Kani described Halos OS as a “safety guardrail” for autonomous driving systems that “will intervene” if AI models are about to make an unsafe decisions.
“We help our partners build a safe architecture that we can decompose down for every function and make sure that if any computer or any sensor fails, the system is still architected to take you to a safe place,” he said.