It is estimated that in 2023, the production of mounted models will exceed 300,000, and the global pre-installation mass production market will exceed USD 23 billion in 2030. Lidar is a standard configuration for high-end autonomous driving, and there are multiple performance evaluation dimensions of explicit parameters, implicit indicators, and measured performance. OEMs put forward the requirements for getting on the car from many aspects, actively participate through investment or cooperation, and tend to customize or self-developed software algorithms. According to calculations, we believe that the output of pre-installed mass-produced models that have been confirmed to be equipped will exceed 300,000 units in 2023, and the price will be concentrated in 400,000-800,000 yuan. In 2024, the global LiDAR pre-installed mass production market will exceed 100 Ten thousand; in 2028, the global pre-installation mass production market will exceed US$10 billion, with a compound growth rate of nearly 90% from 2021 to 2030, and an overall pre-installation penetration rate of 45%.
Numerous technical routes have driven cost reduction and efficiency enhancement, and the development of solid-state, chip-based, and intelligent development has passed the cost and vehicle regulations. Lidar has different routes in five aspects: ranging principle, laser emission, laser reception, beam manipulation and information processing. Innovative technologies can be combined to improve performance and cost. Ranging principle: FMCW program innovation, long-term coexistence with the time-of-flight method. Laser emission: VCSEL emitters promote mass production and cost reduction, and 905nm and 1550nm light sources may coexist. Laser reception: the mainstream use of APD, SPAD or SiPM instead of it has become a consensus. Beam manipulation: The mechanical maturity is the highest, and the ASP has been significantly reduced in recent years; the hybrid solid-state is the fastest to get on the car and it is a consensus, and the MEMS and rotating mirror solutions are about to increase; the solid-state maturity is low and it is expected to become the mainstream in the long term. Information processing: The standard configuration of the main control chip is FPGA, which can coexist with SoC for a long time.
The upstream of the industry chain is monopolized by overseas optoelectronic giants, and lidar manufacturers have self-developed barriers. The upstream market will reach 11.2 billion US dollars in 2030. The three core components of lidar are laser transmitters, photodetectors, and beam manipulation components. They are mainly monopolized by overseas optoelectronic giants such as Lumentum, Hamamatsu, and AMS. Domestic substitution is just starting. The core control points of the innovative technology route are different. Lidar manufacturers mostly use internal research and external expansion to create barriers; in the long run, innovative technologies need to be integrated in the industry, and manufacturers can authorize upstream suppliers to OEM core components to standardize in various ways Products, expand scale and reduce costs.
Recently, leading players have been listed one after another, and mass production is about to roll out in 2021. In 2020, the wave of lidar listings will start, and manufacturers mostly go public through SPAC; the financing investment mostly lies in self-built factories to reduce costs and increase efficiency through large-scale production. The focus has shifted from the autonomous driving market to the pre-installed market, and different manufacturers have different positioning and strategies. In the process of evaluating lidar manufacturers, net profit margin, growth rate, investment efficiency, and risk are the four major factors we pay attention to; indicators such as performance, cost, volume, production capacity, car certification, and car company orders help us to track The development of the manufacturer makes a judgment on the value of the element.
The development of the intelligent driving industry has fallen short of expectations; the commercialization process has fallen short of expectations; supporting policies have fallen short of expectations; the technology has matured less than expected; the cost reduction has fallen short of expectations.
1 Lidar is necessary for high-end autonomous driving, and the first year of mass production of pre-installation is starting
"Intelligence" is the core keyword and main line of our investment in the era of smart cars, and the smart driving system is the core incremental part of smart cars that is different from traditional cars. According to functions, it can be divided into three levels: perception-decision-execution.
At present, the perception layer is mainly divided into two groups: 1) The vision-led group that focuses on camera + millimeter-wave radar and focuses on artificial intelligence vision algorithms, represented by Tesla (TSLA.US) (visual pioneer Mobileye has invested in lidar Research and development); 2) The lidar school with lidar as the mainstay, millimeter-wave radar, and camera as a supplement, represented by Waymo and Baidu (BIDU.US).
L3 is the watershed of autonomous driving, which represents the transformation of initiative from people to cars. There are still issues such as supervision and consumer education; when automakers launch models with L3 functions, they still tend to be positioned as L2.5-L3 level. As a "what you see is what you get" sensor, lidar can enhance the redundancy of the sensing system, supplement the missing scenes of millimeter wave radar and cameras, and cooperate with high-precision maps to play a positioning role. In L3 and above autonomous driving systems, the role of lidar will shift from assistance to dominant, and the number of equipment will also increase.