Zhixin Finance APP has learned that Dongfang Securities has released a research report stating that Robotaxi promotes the iteration of autonomous driving technology. In 2016, Tesla (TSLA.US) turned to self-developing autonomous driving software and hardware, and in May 2023, Tesla FSD V12 became the world's first end-to-end autonomous driving system. While entering L4 level autonomous driving, Baidu (09888, BIDU.US) uses a dual-line parallel research and development strategy to lay out L2+ autonomous driving, providing Apollo autonomous driving technology solutions to complete vehicle manufacturers. In addition, it is expected that Robotaxi will gradually achieve commercialization, partially replacing B-side taxis and online car-hailing, and will accelerate the process of automobile intelligence.

Analysis of the Robotaxi industry chain. The upstream of the Robotaxi (self-driving taxi) industry chain includes various autonomous driving software and hardware suppliers, the middle stream includes autonomous driving solution suppliers and vehicle manufacturers, and the autonomous driving solution suppliers integrate sensors, domain controllers, software algorithms and other hardware and software into a complete L4-level autonomous driving solution and deploy it to the whole vehicle. The downstream of the industry chain is the self-driving travel service platform.

Analysis of Robotaxi's autonomous driving technology route. In 2016, Tesla turned to self-developing autonomous driving software and hardware, and in May 2023, Tesla FSD V12 became the world's first end-to-end autonomous driving system. The end-to-end (End-to-End) model uses a unified neural network model. The vehicle inputs the original data collected by the sensor into the neural network. After processing, the neural network directly outputs the driving command of the autonomous driving vehicle. With the continuous iteration of Tesla FSD V12, it is expected that Tesla Robotaxi will become the first case to apply the end-to-end system to the L4 scene. While entering L4 level autonomous driving, Baidu uses a dual-line parallel research and development strategy to lay out L2+ autonomous driving, providing Apollo autonomous driving technology solutions to complete vehicle manufacturers. The Baidu Apollo autonomous driving has formed a full-stack vehicle intelligence product matrix of “driving, cabin, and map”, and has restructured the large model autonomous driving technology stack. The Baidu L4/L2+ technology collaboration has currently achieved the unification of technology architecture, visual perception solutions, and maps, and the two have opened up data and share autonomous driving infrastructure.

Analysis of Robotaxi's profit cycle. From the perspective of the whole life cycle, Robotaxi's operating costs mainly include: vehicle hardware costs, safety operation costs, and operational costs. Among them, intelligent driving software costs, cloud platform construction costs, etc. can be regarded as fixed costs that are invested once, and are expected to be gradually diluted as Robotaxi expands. At present, more attention should be paid to whether the UE model of Robotaxi can be positive, and when a single car can recover its cost. Under certain assumptions, Dongfang Securities estimates that the RT6 single-vehicle payback period of Carrot Racing is about 2.81 years.

Robotaxi competition pattern: Strong autonomous driving companies and vehicle manufacturers are expected to gain a certain market share. At present, the following Robotaxi commercial cooperation models have been formed: independent operation of vehicle manufacturers, independent operation of autonomous driving technology companies, and cooperation among autonomous driving technology companies, vehicle manufacturers, and travel service operators. In addition to the above three modes, some vehicle manufacturers with strong intelligent driving capabilities and platform advantages may also become the main operators of Robotaxi. At present, Baidu, Xiaoma Zhixin, and others are leading in Robotaxi operations, but it is expected that strong autonomous driving companies and vehicle manufacturers will be able to occupy a certain market share in the Robotaxi market as the Robotaxi trial is gradually rolled out.

Investment recommendations and targets.

Vehicle recommendations: Changan Automobile, BYD, Jianghuai Automobile, SAIC Group; Parts recommendations: Bethel, JWHR, Baolong Technology, Foryou Corporation, Huizhou Desay SV Automotive, Asia Vets, Keboda Technology, and more.

Risk warning

The impact of macroeconomic downturn on automobile demand, the slower-than-expected development of autonomous driving technology, the slower-than-expected commercialization process of Robotaxi, the slower-than-expected promotion of autonomous driving policies, and changes in assumptions affecting the calculation results.