公司于2017 年推出人工智能感知软件HyperVision，其将激光雷达和摄像头收集的原始传感器数据转换成为可以被自动驾驶汽车和机器人直接使用的高质量感知信息。随后根据自推出HyperVision 以来积累的经验，开发出多传感器融合感知软件。公司结合激光雷达硬件产品和其他传感器，形成全栈式车载感知信息采集与处理解决方案——“超级传感器”。公司的全栈一体化软硬件赋予了公司相比于仅仅提供激光雷达硬件产品的供应商更加灵活的商业化机会。Report summary:
The company's gross margin is under pressure in the short term, but we think profitability will recover as future shipments increase. Judging from the industry development logic, the increase in the penetration rate of automotive lidar requires a reduction in the cost of lidar as a prerequisite. Only when the price of lidars sold by suppliers to car companies falls can car companies be more willing to install lidars on lower-cost models. For lidar suppliers such as Sagiteng Juchuang, falling lidar prices will indeed affect gross profit margins in the short term, but on the one hand, companies can reduce costs by continuously optimizing technical solutions for products, and on the other hand, costs can be diluted as product volume brings sufficient scale effects. The combination of the two factors is expected to cause gross margin to rise to a reasonable steady level.
The company's current lidars are divided into R, M, E, F platform products, etc., and the product matrix is constantly being enriched. R platform products are traditional mechanical lidars, mainly used in robotics and other non-automotive industries. M and E platform products are currently automotive-grade solid-state lidars mainly used for ADAS applications. M platform products are the main lidars installed in vehicles. E platform products are specially designed for short-range detection and installed in addition to the main lidar used for blind spot detection. They are compatible with M platform products to provide zero blind spots for automobiles. F platform lidar products are still under development and will support an ultra-long detection range. They are expected to be used in commercial vehicles, railway transportation and other scenarios requiring a long detection distance.
The company launched the artificial intelligence sensing software HyperVision in 2017, which converts raw sensor data collected by lidars and cameras into high-quality perception information that can be directly used by autonomous vehicles and robots. Multi-sensor fusion sensing software was then developed based on the experience accumulated since the launch of HyperVision. The company combines lidar hardware products with other sensors to form a full-stack automotive sensing information acquisition and processing solution - “super sensor”. The company's full-stack integrated hardware and software gave the company more flexible commercialization opportunities compared to suppliers that only provided lidar hardware products.
Customer resources in the automotive industry are high quality, and sales are rising steadily. The company's main customers include Geely Automobile, GAC Aian, Great Wall Motor, Xiaopeng Motors, Lotus, Lucid, etc. The company sold about 7,200 lidar units in 2020, about 16,300 units in 2021, and about 57,000 lidar products in 2022. According to the company's latest announcement, lidar sales reached about 256,000 units in 2023, showing a steady upward trend.
Profit forecast and rating: The company's net profit for 2023-2025 is estimated to be -1,320 million yuan, -1,307 billion yuan, and -1,128 billion yuan, respectively, and EPS of -2.94 yuan, -2.91 yuan, and -2.51 yuan, respectively. Covered for the first time, a “gain” rating was given.
Risk warning: The passenger car market falls short of expectations, and product development falls short of expectations.
