Gasgoo Munich- He Xiaopeng, chairman of XPENG, recently sparked industry debate by suggesting that LiDAR is no longer a necessity in the automotive sector. What role will the technology actually play in China's 2026 auto market? The answers lie in market data, regulatory shifts, and the technical practices of various automakers.
LiDAR Moves Toward Mass Adoption
From a market scale perspective, LiDAR penetration in China's passenger vehicle market continues to climb. Data from Gasgoo Auto Institute shows that total installations of main LiDARs reached 2.756 million units in 2025; Hesai Technology contributed 1.143 million units, securing the top spot. The institute notes that penetration in new energy passenger vehicles hit 21% in 2025, crossing the 16% "chasm threshold"—signaling that LiDAR has officially entered a new stage of mass adoption.
The market expanded further in 2026. In the first quarter, domestic passenger vehicle LiDAR installations exceeded 985,000 units. Hesai led with a 34.9% share, followed closely by Huawei Technologies at 32.3%. Together, they command 67.2% of the market, creating a duopoly. Innovusion and RoboSense form the second tier with 16.8% and 11.0%, respectively.
Hesai's dominance is particularly striking. By March 2026, it held a 55% share of China's main LiDAR market—surpassing all other competitors combined—and has ranked first for 14 consecutive months. Globally, Yole Group's "2026 Global Automotive Market Report" places Hesai first in ADAS main LiDAR shipments with a 43% market share.
On the cost front, prices have plummeted. Data indicates that in 2025, the price of a main LiDAR for L2 driver assistance dropped to around $200, down from $500 two years ago. Chip integration has been the primary driver of these savings.
Image Source: Leapmotor
Falling costs are trickling down to vehicle configurations. Gasgoo Auto Institute previously noted that by 2025, LiDAR was available on models priced as low as 150,000 yuan. Now, with the launch of models like the Leapmotor A10 and the Changan Qiyuan Q05 Laser Extreme Smart Edition, LiDAR has penetrated the 80,000 yuan segment—a key step toward democratizing intelligent driving.
Safety Baselines and Differentiated Choices
Regulators made significant strides toward L3 autonomous driving in 2026. From January 1, mandatory national standards—including the "Technical Requirements for Automobile Vehicle Information Security" and "General Technical Requirements for Automobile Software Upgrade"—took effect. Then, in December 2025, the Ministry of Industry and Information Technology (MIIT) issued the first batch of L3 autonomous driving access permits.
Notably, LiDAR became a mainstream core safety feature among the first batch of L3-approved models. The Arcfox Alpha S, for instance, comes standard with multiple LiDARs to build perception redundancy. However, models like the Changan Shenlan SL03 passed inspection using a pure vision solution. Regulations do not mandate specific sensors.
Industry analysts suggest that at the L2 level and below, LiDAR's core role is "safety redundancy." But at L3 and L4, it shifts from a backup system to a core perception component within a multi-sensor fusion framework, requiring simultaneous decision-making and comprehensive judgment across systems.
Image Source: XPENG
In 2026, the market shows a diverse landscape of technical routes. The pure vision approach, championed by XPENG and Tesla, continues to advance. On May 21, Tesla announced that its supervised FSD is available in China, relying on a pure vision system. XPENG is also sticking to pure vision; its first mass-produced Robotaxi rolled off the line in May 2026, powered by a pure vision setup and the second-generation VLA large model.
He Xiaopeng explained that the effectiveness of XPENG's LiDAR-free solution hinges on the second-generation VLA model's performance in extreme weather, darkness, and complex conditions. He argued that high computing power, large models, and high-density sensors can achieve strong results. Liu Xianming, head of XPENG's General Intelligence Center, added that the necessity of LiDAR depends on the technical route; for users, what matters is the actual performance of autonomous driving, not the sensor specs.
Image Source: Huawei Qiankun Intelligent Automotive Solutions
Meanwhile, the multi-sensor fusion route remains the market mainstream in 2026. Huawei's LiDAR technology continues to iterate, with its 896-line dual-optical-path image-grade LiDAR now equipped on multiple models. The Voyah Tai Shan Ultra Edition, for example, features a four-LiDAR matrix and 34 perception sensors in total. Top-tier players like Li Auto and NIO are also doubling down on LiDAR. Industry competition is shifting from "how many LiDARs are installed" to system robustness and safety in extreme scenarios.
Conclusion
Taken together, the parallel existence of pure vision and multi-sensor fusion will persist in China's 2026 autonomous driving market. Gasgoo Auto Institute forecasts that LiDAR installations and penetration rates will keep rising, unlocking further demand.
With costs falling, regulations clarifying, and leading automakers largely adopting fusion, that path remains critical. Yet the pure vision route is also evolving. Both approaches have their own logic and use cases; their healthy competition and coexistence are driving the continuous advancement and deployment of intelligent driving technology.
