Previously, we explored whether LiDAR constituted a deviation in the autonomous driving landscape, given the industry's embrace of pure vision solutions. It was anticipated that LiDAR would fade from autonomous vehicles due to prevailing trends. However, on March 11, Hesai Technology unveiled its Q4 2024 and full-year financial reports, revealing an annual revenue of 2.08 billion yuan in 2024, with Q4 revenue surging to 720 million yuan. Total LiDAR shipments for the year amounted to 501,889 units, a remarkable 126.0% year-on-year increase; Q4 shipments reached 222,054 units, marking a significant 153.1% year-on-year jump.

On March 12, Hesai Technology announced a groundbreaking multi-year exclusive partnership with a leading European OEM to supply high-performance ultra-long-range LiDAR for its next-generation vehicle platforms, encompassing multiple fuel and new energy vehicle models. This long-term collaboration extends to 2030 and represents Hesai's largest order to date in the realm of overseas factory-installed mass-produced LiDAR. This trend underscores LiDAR's enduring popularity in autonomous driving. In fact, BYD's "Tian Shen Zhi Yan" (God of Heaven's Eye) intelligent driving system also incorporates LiDAR in its premium variant, demonstrating that the industry cannot readily abandon LiDAR in the pursuit of advanced autonomous driving in the near term. LiDAR finds itself both rejected and sought after.

LiDAR's Journey from Skepticism to Popularity

Early autonomous driving technology solutions met with skepticism regarding the use of LiDAR. Prices ranging from tens of thousands to even millions of yuan cast doubt on the imminent commercialization of autonomous driving. The reliability and lifespan of products at such high prices were uncertain, and software algorithms and ecosystems were not fully mature. Industry pioneers like Elon Musk publicly questioned the necessity of LiDAR, asserting that pure vision solutions were the future of autonomous driving. The exceptional performance of Tesla's FSD further validated the feasibility of pure vision. Despite skepticism, many companies chose LiDAR due to its irreplaceable advantages, and others continued to innovate in LiDAR technology, transitioning from initially high costs and technical barriers to subsequent cost reductions and scale expansion.

LiDAR employs the principle of active laser emission and reflection to collect real-time three-dimensional point cloud data of the surrounding environment, enabling centimeter-level distance measurement and precise modeling. In recent years, continuous advancements in optical, mechanical, and signal processing technologies have significantly enhanced LiDAR's scanning speed, ranging accuracy, and environmental adaptability. The introduction of multi-beam scanning technology and solid-state, semi-solid-state technology has allowed LiDAR to achieve multi-angle environmental perception with no dead zones. Through high-speed rotation or integrated design, LiDAR not only significantly improves data acquisition efficiency but also enhances vibration and interference resistance. Advanced filtering, data correction, and deep learning algorithms ensure stable and accurate data output by LiDAR in harsh environments such as rain, fog, and dust. These technological breakthroughs have not only significantly improved LiDAR's practical performance but also laid a solid foundation for multi-sensor data fusion.

Domestic LiDAR companies like Hesai and RoboSense have successively launched LiDAR products for passenger vehicle ADAS since 2021. After several years of exploration, LiDAR technology has achieved breakthroughs, and prices have plummeted, with some products now priced below 10,000 yuan. This transformation has gradually shifted LiDAR from a luxury feature to a mainstream one. New energy vehicle manufacturers such as NIO, XPeng, and Li Auto have incorporated LiDAR into their mid- to high-end models in recent years, enabling advanced autonomous driving functions like city NOA and highway NOA. It can be said that LiDAR's journey from unaffordability to popularity is a tale of self-innovation and continuous validation of its value amidst skepticism.

Does Sensor Fusion Make LiDAR More Affordable?

In autonomous driving systems, no single sensor can meet the comprehensive demand for environmental perception. LiDAR, cameras, millimeter-wave radars, and ultrasonic sensors each have their unique strengths, and only through multi-sensor fusion can a safe, redundant, and stable perception system be constructed. Cameras excel in color and semantic information recognition but lack depth ranging and long-distance detection capabilities. Millimeter-wave radars perform well in penetration but lack the resolution and accuracy of LiDAR. By actively emitting infrared light signals, LiDAR remains unaffected by external light and maintains high-precision ranging in high-speed and nighttime driving environments. Multi-sensor fusion technology leverages the complementary advantages of each sensor to enable vehicles to safely achieve advanced autonomous driving.

Today, LiDAR is considered a standard feature of advanced autonomous driving systems, serving as a crucial safety guard. Consumer perception of LiDAR has evolved from viewing it as a symbol of high technology to recognizing it as a core component of active safety. As LiDAR technology matures, its application scenarios continue to expand, with more automakers attempting to equip mid- and low-end models with LiDAR. The ZERO 1 B10, which went on presale on March 10, is equipped with LiDAR at a price of 129,800 yuan and enables city NOA functionality. This undoubtedly makes intelligent driving accessible to a broader audience, promoting the popularization of autonomous driving technology and providing automakers with new competitive advantages in a fiercely competitive market.

What is the Future of LiDAR?

Currently, autonomous driving technology is undergoing a rapid transition from Level 2 to Level 3 and beyond. At Level 2, vehicles rely primarily on driver assistance systems for partial automation, but drivers still need to intervene in complex scenarios. Level 3 systems require vehicles to possess a higher level of perception, decision-making, and execution capabilities, relying on multi-sensor redundancy to ensure safety. Due to its high precision, all-weather, and long-range ranging capabilities, LiDAR is poised to become an indispensable key hardware component in Level 3 autonomous driving systems. Only through the integration of LiDAR with cameras, millimeter-wave radars, and other sensors can sufficient data redundancy and real-time warnings be provided in emergency situations, thereby ensuring vehicle safety in fully autonomous driving mode.

Beyond autonomous vehicles, embodied intelligence, low-speed logistics and delivery, smart cities, and other fields also have significant demand for LiDAR. With the proliferation of industrial and service robots, the requirements for environmental perception and precise positioning are increasing, and high-performance LiDAR is becoming a core sensor in these areas. Hesai Technology's JT series of robot-specific LiDAR, with its high-precision 3D scanning and real-time data processing capabilities, provides robust support for robot navigation and obstacle avoidance in complex environments. As the application scenarios of embodied intelligence continue to expand, LiDAR deliveries in this market may experience explosive growth, further boosting corporate profitability and creating hundreds of billions of yuan in commercial value for the global smart terminal market.

From the current development trend, LiDAR stands at a pivotal historical juncture. Industry participants are continuously breaking technical barriers, reducing costs, and expanding production capacity to transform LiDAR from a high-end specialized device into a popular and standardized safety core component. Consumers' awareness of LiDAR's safety value is gradually increasing, evolving from initial skepticism and a wait-and-see attitude to calm acceptance. As the market for autonomous driving and embodied intelligence continues to expand, LiDAR is poised to unlock a new blue ocean market worth hundreds of billions of yuan.

Returning to the initial question: Is LiDAR a detour in autonomous driving? At the forefront of intelligent driving, we believe that technology development is iterative. While pure vision solutions may exhibit superior performance in the future, LiDAR remains indispensable for achieving advanced autonomous driving at this stage. In 2024, there was considerable buzz around pure vision solutions, but to date, many automakers' advanced autonomous driving systems still rely on LiDAR. Tesla, emblematic of pure vision solutions, has updated its city NOA function in China, and numerous evaluations have been conducted. However, its performance has not been as "intelligent" as anticipated, leading some to joke about its "acclimatization issues." In contrast, many domestic automakers' advanced autonomous driving systems equipped with LiDAR have demonstrated exceptional performance in city NOA functionality, further highlighting the importance of LiDAR for autonomous driving at this stage. What do you think the future holds for LiDAR? We welcome your comments and invite you to join the discussion!

-- END --