Wednesday, March 18, 2026

NVIDIA partners with BYD, Geely to push for autonomous driving


(China Daily) NVIDIA is partnering with two of China's biggest automakers, BYD and Geely, to tap into the growing autonomous vehicle market as the chipmaker looks beyond artificial intelligence for growth.  

NVIDIA founder and CEO Jensen Huang made the announcement on Monday during his keynote address at the company's annual GTC conference in San Jose, California, declaring that "the ChatGPT moment of self-driving cars has arrived".  

The two Chinese automakers will deploy NVIDIA's Drive Hyperion autonomous vehicle development platform, an integrated system combining chips, computers, sensors and software engineered specifically for the development of Level 4 autonomous vehicles.  

Level 4 vehicles are capable of operating without human intervention within predefined areas or circumstances.  

Japanese automakers Isuzu and Nissan were also added to NVIDIA's robotaxi platform as part of the same announcement.  

The Hyperion platform is designed to support the full lifecycle of autonomous driving development, from the AI models trained in the cloud to the systems that process decisions on the road.  

The autonomous vehicle business itself is much larger than people think, as every autonomous vehicle company in the world is working on some kind of autonomous vehicle, Huang told reporters at a press conference on Tuesday.  

"It's only limited, because driving is limited by butts on seats. If we were not limited by butts on seats, the number of miles driven in the world will definitely go up," he said, adding that when road travel becomes largely autonomous, it would be a "multitrillion-dollar business".  

Autonomous vehicles designed to transport passengers on demand without a human driver, commonly known as robotaxis, represent one of the most commercially significant applications of Level 4 technology.  

While most consumer vehicles currently on the market remain at Level 2, which require drivers to continuously monitor the road, a growing number of companies are deploying Level 4 robotaxi fleets in select cities.  

WeRide, a Chinese autonomous driving technology company, displayed its Robotaxi GXR at the GTC conference. The vehicle was built with the NVIDIA Drive Hyperion platform, which WeRide said helps reduce system costs while accelerating safe and reliable Level 4 robotaxi operations, and "easier cross-market validation".  

The company has outlined ambitious expansion plans, targeting a fleet of more than 2,600 robotaxis in operation globally this year, and tens of thousands globally by 2030.

"In China, BYD and Geely and XPENG and Li Auto, they're all our partners and customers. They're doing great, and they're going to continue to do great," said Huang at the press conference.  

"We standardized on a platform architecture — sensors and computing with all of them, called Hyperion. So when their car goes to Europe, maybe some countries who are unable to accept their software stack, the NVIDIA software stack can be used," Huang explained.  

Monday's announcements add BYD and Geely to a roster of Chinese electric vehicle makers already working with NVIDIA's chips for intelligent driving applications. The list includes Hyper, the premium EV brand under GAC AION, as well as XPENG, Li Auto and ZEEKR.  

China's autonomous vehicle market is expanding at a rapid pace, with cities including Beijing, Shanghai and Shenzhen actively authorizing driverless operations in designated urban zones. Domestic players such as Baidu and Pony.ai are already running fully driverless commercial services in those cities.  

That scale makes China a valuable environment for refining AI-powered driving technology, said industry experts. Autonomous driving is fundamentally a data problem, as the more miles driven, and the more cases outside of normal operations that a system encounters, the stronger its underlying AI becomes.  

By deploying its platform across millions of vehicles in China, NVIDIA could generate the real-world training data needed to outpace competitors and accelerate improvements to its platform, analysts noted.  

The latest collaborations come against the backdrop of trade and technology tensions between the United States and China. NVIDIA's AI chips have been a focal point of bilateral negotiations, with successive rounds of US export controls targeting the most powerful semiconductor products.  

However, automotive applications have so far largely avoided the most stringent restrictions. The Trump administration recently approved the sale of NVIDIA's H200 chips to Chinese companies, which industry observers viewed as a signal of continued engagement between the two countries in the technology sector.

Source: By Lia Zhu in San Francisco | chinadaily.com.cn | Updated: 2026-03-18 10:29

No comments:

Post a Comment