To Find China’s Best Driverless Technology, Look in Silicon Valley
China’s homegrown search giant, much like its U.S. counterpart, has a division focused entirely on driverless vehicles. And just like its rival, Google-born Waymo, both efforts are based in Silicon Valley.
It’s not only Baidu with a toehold in Northern California. China’s self-driving startups are sprouting major R&D outposts 6,000 miles from Beijing. China’s congested megacities may have a need for self-driving cars, but the expertise is elsewhere.
Just ask Pony.ai founders James Peng and Tiancheng Lou. When they decided over a drink in Beijing three years ago that it was time to leave Baidu’s self-driving car unit, the plan was always to start in California. “Silicon Valley is definitely the place to be,” Peng said in an interview. “That’s where all the talent is. China has a lot of raw talent, but with hardcore artificial intelligence, it takes years to build up. China has work to do.”
Baidu recognized this fact back in 2014 when it created an office in Sunnyvale, California, with 200 engineers and scientists working on A.I. and self-driving software. The tech giant said it’s currently adding 150 more people in a second facility in the same town. That research hub has been an overseas incubator for Chinese programmers and engineers, many of whom gain skills from U.S. institutions and colleagues before returning to China. But some leave Baidu and launch startups such as Pony and two other Chinese-backed self-driving firms, Roadstar.ai and Jingchi.ai, while poaching from elite U.S. tech companies.
The push into self-driving cars is part of China President Xi’s mission to make dominance in artificial intelligence a pillar of the country’s economic future. To get there, however, many of China’s biggest tech giants and most interesting startups need Silicon Valley’s knowledge. Engineers working in the U.S.—some Chinese visa holders who went overseas for degrees, some American hires—are a major part of China’s push to develop autonomous vehicles. Didi Chuxing, the ride-hailing giant that pushed Uber out of China, also keeps most of its driverless research in the U.S.
New tariffs imposed by the U.S. have come with a threat to block any company with 25 percent Chinese ownership from buying any American company that makes “ industrially significant technology.” But the trans-Pacific nature of these driverless efforts makes President Donald Trump’s push to wall off American technology difficult, if not impossible.
Trump has also made intellectual property a trade issue, and U.S companies worry about theft. In July, U.S. federal prosecutors accused former Apple engineer Zhang Xiaolang of downloading files containing proprietary information on driverless cars and selling it to a Chinese startup.
Regardless of Trump’s concerns, the Chinese tech companies are already here. Even if the Trump administration succeeds in stopping deals, it won’t prevent Chinese companies from hiring data scientists, software engineers and technical leaders from the likes of Alphabet Inc., Tesla Inc., and other major players in the driverless field. Qing Lu, the chief financial officer of Jingchi, came from Velodyne Lidar. Liang Heng, Roadstar’s chief technical officer, worked at Google and Tesla.
Peng started Pony two years ago with about 20 people in Fremont, near Tesla’s car factory, and now has more than 100 engineers working on cars that drive themselves. He has raised $214 million from investors in both the U.S. and China, with the latest being led by China-based ClearVue Partners and Eight Roads Ventures, an offshoot of Fidelity Management.
The company’s test cars can navigate traffic and drive at highway speeds. Those cars have been on public roads thanks to superior testing conditions found in the U.S. China until very recently had a strict ban on driverless test vehicles. Next year, Peng said, Pony will deploy a test fleet of robotaxis in Guangzhou—about the same time General Motors plans to start a driverless ride-hailing business, most likely in San Francisco. Pony plans to deploy the technology commercially, in either ride-hailing or delivery fleets, within three to five years.
Peng exemplifies the way China’s tech leaders come into their own through contact with U.S. institutions and employers. After an education in both China and the U.S., he worked for seven years at Google as a software engineer developing big data applications and artificial intelligence. He was lured away in 2012 by Baidu, where he led its autonomous-vehicle project in Sunnyvale. That’s where he met Pony co-founder Lou, who worked to develop self-driving cars at Google before the formation of the Waymo autonomous unit.
“I was always fascinated by cars and mechanical things,” Peng said. “The potential really fascinated me.”
Roadstar had a similar path. Founder Xianqiao Tong is a Chinese engineer who got started in autonomy by working on Nvidia’s driver assistance systems and then at Apple before joining Baidu.
To lure more talent, Tong set up a large R&D center in Cupertino, where Apple is based, and gave that office responsibility for self-driving tech. Roadstar is targeting 2020 to deploy thousands of robotaxis in Guangzhou, Tong said in an interview. The cars will be able to drive in Level 4, which means they will be fully autonomous within a mapped-out area.
“The Chinese startups are spending money to hire talent and they are attracting top talent,” said Grayson Brulte, co-founder of autonomous vehicle consulting firm Brulte & Co. “By setting up in Silicon Valley, they are getting the knowledge that they don’t have and they can export it back home.”
Roadstar is working with Renovo.auto, a Silicon Valley company that creates the operating system for self-driving cars, to connect its software to lidar, cameras and other sensors, said Renovo CEO Christopher Heiser. The operating system takes what the hardware sees and sends it to Roadstar’s computer brain, which then drives the car. Renovo can also connect Roadstar’s cars to mapping systems and even to a ride-hailing system like Uber or Lyft.
The idea is that even if Roadstar is starting out far behind American tech leaders such as GM’s Cruise or Waymo, the partnership with Renovo could speed up deployment. Roadstar, Heiser said, “learned a great deal working on AV tech in the Valley and have now built an international company that combines Chinese and Silicon Valley features.”
Will the Chinese startups succeed? Several new Chinese automakers made a similar play in electric vehicles in recent years, with mixed results. Faraday Future, based in Los Angeles and partly owned by indebted Chinese billionaire Yueting Jia, has delayed plans to build a plant amid financial trouble and recently started retrofitting an old tire plant. Karma, which was acquired by a Chinese industrialist, is building a luxury plug-in hybrid but sells cars in small numbers.
Earlier this month, however, Chinese-backed EV startup NIO kicked off a $1 billion initial public offering in the U.S. NIO’s chief financial officer, Louis Hsieh, said the company will be in Silicon Valley to develop autonomous vehicles. “There’s not the talent currently in China for autonomous driving,” Hsieh said. “You need lidar for Level 4 autonomy. Most of the talent in the area is in Silicon Valley.”
Thanks to tax credits from the government from the government, China’s automakers have become competitive players in the EV game. Now those same companies, alongside China’s overseas tech firms, are fishing in Silicon Valley’s limited pool for self-driving talent. That will pressure U.S. automakers like GM and Ford Motor.
“With the Chinese entering the market for A.I. and autonomy skills,” said Eric Noble, president of consulting firm The CarLab in Orange, California, “the automakers are having to make exceptions to their normally low pay for engineers. They have to pay up.”