Investment opportunities in automotive electronics brought about by intelligent transformation

This part is easily overlooked. The purpose of the camera and lidar is to sense the external environment, so they are installed on the outside of the vehicle and are clearly visible. The integrated navigation system is generally installed at the center of the rear axle of the vehicle due to its inherent attribute: positioning. It cannot be seen without disassembling the vehicle. From the perspective of the current technical route of intelligent driving development, it is natural for high-precision positioning to become the standard configuration of L3 and above models. Take a look at the picture:
 

The picture above is the most commonly used combination to achieve high-precision positioning: GPS/Beidou dual-mode satellite navigation system + RTK + IMU.

Generally speaking, companies in this link will package and sell a series of soft and hard modules used after the Qianxun differential signal as a combined system, which is what securities firms often call a "position unit."

According to CITIC Securities' estimates, the domestic inertial navigation market size should be more than 50 billion, and the current domestic companies engaged in this field include: Huace Navigation, Daoyuan Electronics, StarNet Yuda, and Zhonghaida.

Huace Navigation: designated supplier for Nezha, Geely Lotus, BYD, and Great Wall;

Zhonghaida: SAIC certification;

StarNet Yunda: Long-term cooperation with Baidu, Meituan, Momenta and Neolithic.

Daoyuan Electronics: Not listed on the stock market, its products are very competitive. It is a high-quality core target second only to Huace in China and is worth paying attention to.

Huace is the absolute leader, and passenger car mass production orders are king. Xingwang Yuda mainly engages in military industry, and its unmanned driving mainly cooperates with the Internet and unmanned logistics vehicles. Zhonghaida still needs to work hard.

Domain controller

When I wrote about the smart driving industry chain before, I didn’t write about domain controllers separately, but when talking about stocks, related companies are often mentioned.

Why do we talk about domain controllers when discussing the investment logic of intelligent automobiles? Because before the trend of intelligence came, the electrical architecture of vehicles was relatively traditional, and most of them were developed with a single ECU as a functional unit, with relatively simple functions. There are often dozens or even hundreds of ECUs on a car. As the degree of automobile intelligence deepens, the number of ECUs required is growing exponentially. For example, the total length of the wiring harness on a car is currently about several thousand meters. If these ECUs are not integrated using a domain controller, the total length of the wiring harness of a single vehicle will increase exponentially with the increase in the number of ECUs in the future, which will undoubtedly greatly increase the difficulty of wiring harness layout. This is just one aspect. Too many ECUs will also lead to complicated control logic, inability to share computing power, and excessive energy consumption. With a domain controller, these problems will be solved.

Although some manufacturers have their own division methods, their functions and division concepts are similar. The industry generally recognizes Bosch's five-domain division method: power domain, chassis domain, smart cockpit domain, autonomous driving domain and body domain. Among these five domains, the smart cockpit domain + autonomous driving domain are the two with the largest share and the highest future growth rate. According to a McKinsey research report, in the next five years, the market size of these two areas will account for more than 40% of the entire automotive domain control.

According to Gasgoo's estimates, by 2025, my country's market shipments of autonomous driving domain controllers will reach 4 million sets, and shipments of smart cockpit domain controllers will reach 5 million sets, with annual compound growth rates exceeding 50%.

When we talk about domain controllers, we will inevitably talk about chips. Currently, the leader in smart driving chips is NVIDIA, while the leader in smart cockpit chips is Qualcomm. The two domestic listed companies I often mention, Desay SV and Thundersoft, are deep partners of these two chip companies: Desay is tied to NVIDIA, and Thundersoft is tied to Qualcomm. This shows the status of domestic smart driving and smart cockpit domain controllers.

Generally speaking: the two leaders, Sinotrans, are relatively hard, while Chuangda is relatively soft; Huayang Intelligent Driving and Cockpit are making progress in parallel; Joyson wants to make money from everything, but it is not so easy to make money from everything; as for Jingwei Hengrun, it can be summed up in five words: Hengrun works very hard.