Hesai held the global launch of Kosmo at its Maxwell super factory in Shanghai on October 14, unveiling a 3D world data engine and disclosing that it has entered robotic motion modules and begun commercial deliveries. Li Yifan said that LiDAR units that sold for more than RMB 300,000 back then can now be produced for around $200, with the production cycle shortened to one unit every five to six seconds.
The article analyzes the computing requirements of vision-only and LiDAR-fusion approaches in autonomous driving, arguing that computing demand is not simply determined by the number of sensors, but by how the perception system processes data and by its algorithm architecture.
Data from NE Times shows that LiDAR installations in Chinese passenger vehicles reached 525,500 units in August 2026. Hesai installed 182,436 units, just 65 fewer than Huawei, and the two together accounted for 69.4%. The growth logic is extending beyond simply adding vehicle models to multiple LiDARs per vehicle, robot perception and 3D spatial mapping.
Why we recommend it: The original paper provides monthly installation and market share data, useful for assessing the automotive LiDAR market landscape
Hesai, Seyond and RoboSense successively released financial reports in August. Comparing the three, the article points out that as mass-produced vehicle models, robotics and embodied intelligence become real demand, the growth logic of leading LiDAR companies is diverging.
RoboSense plans to sell its embodied robotics complete-machine business for RMB 59.878 million and focus on supplying key components to robot manufacturers. Seyond has signed a design win with Chuneng Automotive for a forward-facing primary LiDAR, while Hesai has also announced new moves in spatial intelligence.
In the first half of 2026, RoboSense shipped 719,200 units, up 169.6% year on year, while Seyond shipped 450,900 units, up 364.8%. The article compares the scale, profitability and second growth curves of the three companies.
LiDAR installations totaled 1.595 million units in 2024 and 3.316 million in 2025; 2.282 million units were installed in the first half of 2026, with Huawei's 896-beam LiDAR becoming the new benchmark.
O-Film and Luming Photonics are deepening their collaboration to scale solid-state LiDAR deployment based on the ADS6311 SPAD chip.
Guangzai Technology, an incubated company of the Hangzhou Institute of Optics and Mechanics, collaborated with the University of Cambridge, Southeast University and others to publish a paper in Sustainable Cities and Society, proposing SCE-Model, a spatiotemporal carbon emission model optimized by machine learning and validated by LiDAR. The model output is highly consistent with coherent differential absorption LiDAR observations (R² = 0.79; ±5 ppm), enabling hourly, block-level carbon emission mapping.