With the release of Tesla Robotaxi, the competition between Chinese and American companies in the field of autonomous driving technology has entered a white hot stage.
On October 11th, Tesla released the Robotaxi - CyberCab. Before that, some media reported that Baidu's self driving travel service platform "Apollo Go" was planning to expand its driverless taxi service to overseas markets, and the first destination might be Hong Kong, China.
Some argue that the once seemingly distant autonomous driving technology is now advancing rapidly with the rapid development of end-to-end technologies, especially in the field of Robotaxi.
In an interview with the Daily Economic News, Zhou Guang, CEO of Yuanrong Qixing, said, "End to end is the most promising path to universal physical AI. Currently, it is only in the 1.0 stage of end-to-end. In the next few years, physical AI will continue to see technological breakthroughs, and Robotaxi may not be that far away
Robotaxi accelerates in China, stalls in the US
As one of the earliest American companies to lay out Robotaxi, it once led in this field and gave birth to five star autonomous driving companies: Zoox, Waymo, Cruise, Argo AI, and Aurora. But with difficulties in commercialization and severe losses, American players faced numerous challenges in the future.
According to public information, Zoox was acquired by Amazon in December 2020 at a valuation of $1.3 billion, which is equivalent to half of its previous valuation. Aurora's stock price plummeted after going public through SPAC in 2020, resulting in a net loss of up to $792 million by 2023. Cruise was suspended in 2023 due to a serious traffic accident and has recently resumed small-scale operations. Although Argo AI was funded by Ford and Volkswagen, it ended up bankrupt due to a lack of substantial progress.
Currently, only Waymo, a subsidiary of Google's parent company Alphabet, operates a fleet of 700 Robotaxis in cities such as San Francisco and Phoenix. The latest news is that its autonomous taxi service in Austin, Texas is about to open to the public.
In contrast to the Chinese market, Baidu Apollo Go, which was established at the same time as Waymo, is developing rapidly. Official data shows that as of July 28, Apollo Go has provided more than 7 million rides in 11 cities across the country, with 100 million kilometers of test and operation mileage. According to the information disclosed previously, by the end of 2024, Apollo Go will achieve balance of revenue and expenditure in Wuhan, and will enter a profitable period in Wuhan in 2025.
In addition, other major players in the Chinese Robotaxi field, such as Komatsu Zhixing, Wenyuan Zhixing, Qixing, and Yuanrong Qixing, are also actively promoting commercialization.
For example, recently Yuanrong Qixing announced plans to establish a European operations center in Germany. The CEO of the company, Zhou Guang, said, "Yuanrong Qixing's plan for autonomous driving to go global includes two paths: one is to assist Chinese car companies in expanding their overseas intelligent driving markets; the other is to negotiate and cooperate with foreign car companies to help them expand their markets
According to incomplete statistics from reporters, more than 50 cities in China, including Shenzhen, Shanghai, Jiangsu, Hangzhou, etc., have introduced pilot demonstration policies for autonomous driving. Meanwhile, companies such as Baidu, AutoX Antu, and Xiaoma Zhixing have conducted commercial trials for Robotaxi in specific areas of cities such as Beijing, Wuhan, and Chongqing.
Pacific Securities predicts that by 2025 and 2030, the market size of Robotaxi in China is expected to exceed 1.18 trillion yuan and 2.93 trillion yuan respectively, making it one of the largest autonomous driving scenarios in the market space.
Two technological routes in the game
At the level of technological roadmap, American companies such as Waymo and Tesla mainly adopt the strategy of "bike intelligence" in autonomous driving technology, focusing on achieving autonomous driving through onboard sensors and algorithms. For example, Tesla adopts a purely visual approach for exploration and proposes the BEV+Transformer architecture when big model technology arrives.
In response, Chinese companies have chosen to adopt multi-sensor fusion routes such as LiDAR, which means that the high cost of LiDAR in the early stages will bring certain pressure to operations. Zhang Shiyue, Senior Industrial Analyst at Toubao Research Institute, stated that from the perspective of technology cost breakdown, the LiDAR solution can provide 3D modeling and high-precision environmental recognition, with relatively high costs. The unit price of each high-performance LiDAR is approximately $1000.
But with the advancement of technology and the expansion of scale, the cost of a domestic laser radar has dropped to around 2000 yuan. The latest first quarter 2024 performance report released by Sagitar Juchuang shows that its ADAS application LiDAR product revenue is about 310 million yuan and sales volume is about 116000 units. Based on this, it is roughly estimated that its average delivery price in the first quarter of this year was about 2666 yuan/unit, which is about 38% lower than the level of nearly 4300 yuan in the same period last year.
Compared with the two technology routes in China and the United States, Zhang Yangrui, Director of CIC Insight Consulting, believes that Tesla's route advantage is the ability to monetize earlier and achieve large-scale data accumulation at lower costs. Domestically, Chinese domestic manufacturers tend to take a hardware led approach, combining map, sensor, and radar technologies to achieve more accurate environmental monitoring and vehicle control.
In addition, the Ministry of Information Technology and five other departments have officially announced the list of pilot cities for the application of "vehicle road cloud integration", as well as the demonstration pilot policies intensively introduced by local governments. Chinese enterprises also pay more attention to vehicle road collaboration technology, and improve the safety and efficiency of autonomous driving through the support of the Internet of Vehicles and intelligent infrastructure.
Some people believe that the competition between China and the United States in the field of autonomous driving is not just a competition between individual enterprises, but also a competition between the entire industry chain and ecosystem. And China's technology route that combines infrastructure construction gives it certain advantages in promoting the application of autonomous driving.
Gradually popularized or by 2026
At present, in the field of automatic driving, players such as Tesla Robotaxi, Google's Waymo, and China's Apollo Go jointly constitute the competition pattern of the industry. However, in order to deploy a large number of unmanned vehicles in an orderly manner and achieve commercial closure, in addition to the technology itself, attention should also be paid to multiple aspects such as carriers and operations.
From publicly available information, the strategies adopted by Silicon Valley intelligent driving players to implement Robotaxi mainly include three types: firstly, the self-developed travel network represented by Zoox, which integrates technology, manufacturing, and operation to form an autonomous driving taxi service system; The second type is self developed autonomous driving technology by ride hailing companies represented by Uber and Lyft; The third type is for top car companies to establish subsidiaries to research and develop technology, such as Argo AI, which has been funded by Ford and Volkswagen.
In terms of domestic players, Robotaxi players can be divided into two categories: one is Internet technology companies such as Apollo Go, Xiaoma Zhixing, Wenyuan Zhixing, etc., which provide technical support to achieve commercialization through cooperation with car companies or travel platforms; Another type is traditional car companies, which collaborate with autonomous driving technology companies to develop related technologies, such as Guangzhou Automobile Group and Wenyuan Zhixing's partnership to build the Roboaxi fleet. In addition, new energy car manufacturers such as Xiaopeng Motors are also laying out Robotaxi.
According to a report by consulting firm Roland Berger, although both China and the United States have chosen to promote technological progress through policy and regulatory relaxation on their development paths, they have also combined the characteristics of their respective Robotaxi commercialization progress in the scale and standards of policy relaxation.
Among them, China strictly follows the development attitude of "step by step" for the commercialization of Robotaxi. After the technical reliability has been verified, the regulatory authority will liberalize the space for technological development as appropriate, feed back the technological iteration, and continue to achieve success in the "spiral" path of progress. For example, the idea of Apollo Go is to run Wuhan's profit model and then replicate it to other cities in batches.
The United States has adopted a more open attitude towards development and a "leapfrog, fast-paced" approach. For example, the Cruise accident was caused by the immediate deployment after the road testing and commercial trial operation stages were verified to be mature.
Zhou Guang, CEO of Yuan Rong Qixing, said: "The biggest challenge for autonomous vehicle to scale up is still the technical level. Current technology cannot cope with all complex scenarios." In his view, to achieve technology iteration, we need to obtain a large number of desensitization high-quality data training models. Only by ensuring that under the premise of safety, we promote human-machine driving, and promote high-order auxiliary driving systems to widely load vehicles, can we generate and collect these data.
According to Frost&Sullivan's data analysis, Robotaxi will gradually become popular when its cost per kilometer will be lower than that of manned ride hailing services, around 2026.