China Securities Co., Ltd.: High-level smart driving accelerates popularization, and the in-vehicle hardware market welcomes new development opportunities.

GMTEight APP learned that, recently, China Securities Co., Ltd. released a research report stating that on February 10, 2025, BYD Company Limited (01211) demonstrated its smart driving system research and commercial achievements at its smart driving conference. Its high-end smart driving system in 2025 will be available in vehicles priced as low as 100,000 yuan and even 70,000 yuan, a strategy that will prompt other companies in the industry to follow suit, leading to the widespread adoption of smart driving across the industry. According to EO Intelligence, the penetration rate of L2+ (highway NOA) is expected to reach 8% by 2024, while L2+ (urban NOA) is currently at 0%, estimated to reach 55% and 25% respectively by 2030. Under the backdrop of equal opportunities in smart driving, the penetration rate of smart driving will rapidly increase. Correspondingly, there will be increased shipments of laser radars, cameras, computing chips, and domain controllers, providing opportunities for CKH HOLDINGS' configuration upgrades and an expected surge in demand for vehicle electronic components. The views of China Securities Co., Ltd. are as follows: Advanced smart driving accelerates popularization, AI large models + hardware cost reduction drive equal opportunities in smart driving In the first half of 2024, the penetration rate of L2 intelligent driving in the automotive market in Shanxi Guoxin Energy Corporation had already reached 50%, with high penetration rates, but the functionality is still being upgraded continuously towards L2+/L3. On February 10, 2025, BYD Company Limited held a smart strategy launch conference, introducing its self-developed "Eye of God" technology matrix to promote the penetration of high-end smart driving into the 70,000 yuan market, aiming to increase the proportion of L2+ models from less than 10% to 30% by 2025. In addition to cost reduction in the supply chain, which allows for "increased quantity without increased price" for vehicles, BYD Company Limited is also using DeepSeek big models to strengthen data mining and algorithm iteration, allowing software capabilities to compensate for hardware simplification. For example, the Eye of God C-level system achieves a dense point cloud effect similar to laser radar through an end-to-end model, reducing the reliance on high-cost sensors. According to EO Intelligence, the penetration rate of L2+ (highway NOA) is expected to reach 8% by 2024, while L2+ (urban NOA) is currently at 0%, estimated to reach 55% and 25% respectively by 2030. Perception: Advanced smart driving brings multi-sensor integration, sensor cost reduction drives rapid demand growth The upgrade of smart driving requires a higher richness in the perception system, and sensor fusion solutions are becoming a trend. Currently, laser radars are being installed in vehicles priced as low as 100,000 yuan, with prices decreasing to the level of thousands of yuan. As the level of autonomous driving gradually increases, the number of vehicles equipped with laser radars is expected to exceed 3, with domestic companies such as Hesai and Hikvision already dominating the market share in the field of laser radars, accounting for 83%. In terms of cameras, the upgrade of smart driving is driving an increase in the number of cameras installed per vehicle, an upgrade in pixel resolution, and an increase in demand for surround-view, in-car, and ADAS cameras, as the equal opportunities in smart driving are prompting an increase in the number of on-board cameras from an average of 4-6 to 11-12, and an upgrade in front view pixel resolution from 2M to 5M, 8M, or even higher. In addition, the quantity of millimeter-wave radars and ultrasonic radars is also increasing. Decision-making: Vehicle computing power enters an arms race, rapid development of domestic smart driving chips Intelligent driving and intelligent cabin domains are major new areas of development for smart vehicles, and the integration of cabins and driving will be an important evolution direction in the future, requiring high-computing power central control chips to support more complex autonomous driving models and cabin integration. Chips in the smart driving domain show a significant head effect, with domestic chip manufacturers actively seizing market share. In the smart driving field, as of 2024, NVIDIA's Orin-X has secured a 39.8% market share in China with 2.1 million installation units, while Tesla's self-developed chips have a 25.1% market share with 1.32 million installation units, and Huawei and Horizon Robotics also have a certain market share. In the cabin domain, Qualcomm dominates with a 70% market share, while other players include AMD, Renesas, Rockchip, Huawei, and Sychip. Domestic chips are rapidly emerging, and under the trend of domestic substitution, they are expected to gain a higher market share in the future. Execution / Transfer: Vehicle data volume surges, PCB and connector demand expands (1) PCB: As the penetration rate of advanced autonomous driving increases, the number of sensors installed per vehicle will see a significant growth, driving an increase in the usage area of PCBs on the one hand; on the other hand, autonomous driving systems often use HDI boards, so the value of PCBs is expected to increase significantly. Intelligent cabins include various electronic systems such as full LCD instruments, large central control systems, in-car entertainment systems, display systems, and streaming rearview mirrors, leading to a substantial increase in the usage of PCBs. (2) Connectors: Due to the restructuring of the vehicle's electrical architecture, connectors need to meet higher performance standards in diverse application scenarios. The surge in vehicle communication data at the terminal requires an increase in real-time transmission requirements, leading to high demand for high-speed connectors. Recommended Focus: Laser radars, cameras, domain control / computing power, execution / transfer, and other related areas. Risk Warning: Pressure to optimize supply chain costs, challenges in coordinating chip supply, adapting to the rapidly evolving technological landscape, compliance management of data applications, and the need for local adaptation of differentiated standards.