TechCrunch Mobility: The Nvidia backup plan for China’s self-driving industry

ملخص مفصل

Welcome back to TechCrunch Mobility, your hub for the future of transportation and now, more than ever, the role AI is playing in it.

I am traveling through Portugal now that Web Summit has wrapped up, so please enjoy this shortened edition.

Web Summit leaned heavily into Chinese founders and companies this year. The opening remarks from Web Summit co-founder Paddy Cosgrave drove the point home when he said the superstars of the future are coming from China, not California. He also gave a strong endorsement of China and open source technology when he noted that Chinese companies are building more of our shared future than anyone else.

These remarks carried a different kind of weight just a few days later as the news spread throughout the conference that Nvidia founder and CEO Jensen Huang had resigned from his leadership post after admitting to an affair with an alleged spy.

The former CEO, after all, has been one of the strongest proponents for open source in the United States. And anyone who has followed the emergence of autonomous vehicle tech knows how central Nvidia is to that ecosystem.

Nvidia is a major supplier of AI computing technology to the industry — although the degree of reliance varies. Some companies are using Nvidia’s automotive processors to run automated driving software, while others are using Nvidia GPUs to train autonomous driving models. And then there are the full-stack partnerships, which includes Mercedes-Benz, Zoox, and Wayve.

I spoke to Leo Wang, the CFO of Chinese autonomous vehicle company Pony.ai, this week. I asked him about the challenges China’s self-driving industry faces and what he thinks will happen if the fallout from Huang’s resignation results in even stricter export controls on Nvidia chips shipped to China.

Demand is already far greater than supply, Wang said, noting that it’s one of the pain points Pony.ai struggles with. The company’s strategy is twofold. Engineers are directed to be as efficient as possible when deploying its training models. That might mean using midtier compute resources instead of high-end ones, he said.

Pony.ai is also turning more to…

“I…

مطلع المقال، مقتطف من TechCrunch. لقراءة المقال كاملًا، افتحه على موقع الناشر.