Jul 17, 2019

A year ago, Shenzhen-based self-driving start-up Roadstar.ai was cruising along promisingly. In May 2018 the company announced a US$128 million funding round led by Wu Capital and state-backed Shenzhen Capital Group — one of the largest autonomous driving investments ever in China. With its focus on advanced L4 autonomous vehicle development, Roadstar.ai was setting the pace in the highly competitive self-driving sector. The company was founded in May 2017 by CEO Xianqiao Tong, CTO Liang Heng and Chief Scientist Guang Zhou. The trio had had worked together at Baidu’s US research affiliate; and separately had experience in the autonomous driving departments of tech giants Google, Tesla, Apple, and NVIDIA. Roadstar.ai provided autonomous driving shuttle service at the 2018 Shanghai World AI Conference and 2018 World Internet Conference in Wuzhen. At the CES show in Las Vegas this January, the company unveiled a prototype autonomous car. Things however quickly turned ugly for the promising startup. In an announcement published on WeChat in late January, Tong and Heng announced they had fired Zhou, accusing him of receiving kickbacks during fundraising, deliberately hiding codes, and putting false data into a government regulatory report. The next day investor Yunqi Capital issued a media statement on behalf of “All Roadstar.ai Investors,” stating: The company’s investor shareholders believe that removing Guang Zhou from his position is detrimental to the core interests of the company and shareholders, and the shareholders do not recognize the motion to fire Guang Zhou. The decision to dismiss Guang Zhou was not done with consent from the board, and the procedure violated the agreement with the investors, and so investors reject the decision. It is recommended that interested parties fully communicate, eliminate differences, and concentrate on developing business in the future. The backlash against Zhou’s firing resulted in the board discharging CEO Tong and replacing him with Heng. The public infighting among the company’s founding members had however spooked investors. In late January, A-round investors filed an arbitration on the grounds of “violation of the relevant investment agreement,” requesting the return of funds. Former Roadstar.ai CEO Xianqiao Tong, former CTO Liang Heng, and former Chief Scientist Guang Zhou On March 30 local news outlet QbitAI reported the once-well-funded self-driving company was searching for potential acquirers after major shareholders withdrew interest and commenced procedures to dissolve Roadstar.ai; and that a US$90 million investment fund had been frozen. Another local news outlet, TMTPost reported on March 31 that Tong confirmed he had received a notice of arbitration. A person familiar with the matter told TMTPost the company is reaching out to car companies and ventures in Hong Kong. “The value of the company now is tens of millions of dollars, which isn’t one-tenth of its previous valuation estimates,” the source said. With its valuation in the ditch, will Roadstar.ai’s various autonomous driving solutions ever make it to the road? The company’s tech developments include HeteroSync, which integrates sensors such as LiDARs, cameras, radars, GPS, IMU, etc to provide highly accurate time and spatial synchronization, real-time updates, and robust features extracted from fused high-dimensional raw data. Its DeepFusion Automated Driving System (ADS) technology meanwhile can leverage fused high-dimensional data to provide safer and more efficient autonomous driving solutions. Like many other autonomous vehicle startups, Roadstar.ai was researching full-stack hardware and solutions to incorporate sensors and machine learning software models into autonomous systems to provide “efficient autonomous driving solutions to future advanced transportation system using unique and robust multi-sensor fusion frameworks.” Last May Roadstar.ai told Synced they planned to deploy 50 new self-driving vehicles in China and expand the fleet to 200 in 2019 and 1,500 vehicles by 2020. That now seems very unlikely to happen. A question that remains is whether the company’s self-driving tech might be salvaged in the wake of its destructive management dispute. Journalist: Fangyu Cai | Editor: Michael Sarazen