With the global transition of autonomous driving from L2 to L3+, in the battle between LiDAR and pure visual perception routes, LiDAR is redefining the industry landscape at an astonishing pace of technological evolution and quietly building a new industrial ecosystem in the era of intelligent travel.
Before the end-to-end model of autonomous driving became mainstream, there were many discussions and controversies about the necessity of configuring LiDAR as a sensor. Some argue that LiDAR is a necessary condition for achieving advanced intelligent driving, especially in terms of efficiency in detecting objects and issuing corresponding commands. However, in 2024, some companies such as Xiaopeng, Ledao, Hongmeng Zhixing, etc. began to test the pure visual route; On the other hand, industry leaders such as Musk have repeatedly criticized LiDAR on social media and in public, believing it to be the wrong solution.
Overall, LiDAR and pure visual path have their own advantages and disadvantages in autonomous driving technology. Lidar performs well in application scenarios that require high-precision environmental perception due to its high precision, independence from lighting conditions, and wide field of view coverage; The pure visual path, with its advantages of low cost, rich data volume, and simple system, is more attractive in the early stage of large-scale commercialization of autonomous driving technology. But with the advancement of intelligent driving technology, the evolution of LiDAR technology, and the cost reduction brought about by economies of scale, LiDAR has quietly entered a golden age of explosive growth.
The 'Necessary Debate' of Lidar
In 2024, intense discussions about LiDAR continue to unfold. The collision of industry leaders' viewpoints reflects the deep game of technology route selection.
At the Global Automotive Electronics Show (CES) in early 2024, Dmitri Dolgov, co CEO of Waymo, stated that LiDAR is the "optical seatbelt" for L4 autonomous driving, and its centimeter level accuracy and multidimensional data stream are the underlying guarantees that pure visual solutions cannot replace. Tesla's head of autonomous driving, Ashok Elluswamy, insists on a "vision first" approach: "Humans can drive with only binocular vision, and LiDAR is just a redundant design during the transition period.
As the leader of the pure visual route, Musk stated on Tesla Investment Day that "from a cost perspective, the pure visual solution is more competitive because the cost of LiDAR is high and difficult to afford". Laser radar is the wrong solution, "Musk reiterated his criticism of laser radar technology on social media more than once, insisting that simulating biological neural networks and the visual system of the eye is the most effective technological path.
As a comparison, in 2024, many domestic occasions, enterprises, industry leaders, and experts have expressed their support for the necessity of LiDAR. At the L6 press conference, Liu Tao, CEO of Zhiji Motors, said, "In this era of intelligence, if you buy a smart electric car worth around 200000 yuan without a LiDAR, it's as strange as buying a smartphone without a camera." On NIO Technology Day, Li Bin said in an interview, "Who says LiDAR is useless? It must be either stupid or bad.
During a media interview, Yu Chengdong delved into the issue of LiDAR. Yu Chengdong said, "We insist on using LiDAR because LiDAR can improve safety. In several cases, the camera may become blind and invisible, while LiDAR does not." Yu Chengdong emphasized the importance of LiDAR in the safety of autonomous driving, believing that pure visual solutions have limitations, especially in extreme or complex environments where they may fail.
Of course, Huawei has not bypassed the pure visual solution. When talking about Huawei's pure visual solution, Yu Chengdong clearly stated, "We have a low-end solution without LiDAR, which is our basic solution, which can reduce costs. However, its security cannot reach that of LiDAR." In terms of technical route selection, Yu Chengdong said that there is only one standard: what can bring a true security experience. There is no such thing as' technological fundamentalism ', nor is there any use of' advanced 'technology. Yu Chengdong bluntly said, "Life is the most precious thing. No matter how expensive something is, compared to life, it is not money.
On September 20, 2024, during a media interview at the launch event of the Extreme Krypton 7X, Lin Jinwen, Vice President of Extreme Krypton, said, "Whether it is a high-end luxury brand that produces pure electricity or extended range, Extreme Krypton has not given up on LiDAR. Lidar does have a high cost, but the extreme safety that can be achieved with algorithm software cannot be achieved with pure vision
From his personal perspective, not using LiDAR is just finding a reason to reduce costs. On September 26th, at the GAC AiON RT press conference, Xiao Yong, Vice General Manager of GAC AiON, stated that GAC AiON cannot ignore the lower limit of intelligent driving safety. Its first consideration is safety. While enriching the upper limit of intelligent driving experience, it is also necessary to raise the safety lower limit of advanced intelligent driving. He believes that LiDAR is still irreplaceable in the current environment.
At a previous TEDA forum, Gu Huinan, the General Manager of GAC Aion New Energy Vehicles Co., Ltd., also expressed a similar view, saying, "We are now discussing whether to use pure vision or LiDAR, or whether to use pure vision combined with ordinary radar without LiDAR. However, it is widely recognized that when it comes to L3 level or above, LiDAR is still necessary for safety with dual redundant systems.
At the same time, in the second live broadcast of Ideal's "2024 Ideal AI Talk", Ideal Auto CEO Li Xiang said, "If Musk has driven a car in China, he will also keep the LiDAR." Li Xiang said that Ideal Auto keeps the LiDAR for safety, especially in the complex road environment in China, where it can provide additional perception capabilities and help achieve higher-level autonomous driving.
The sharp drop in cost curve opens up the era of mass production of front loading
Behind this debate is the transformation of LiDAR technology from a "noble accessory" to a "civilian core sensor". Li Deyi, academician of the CAE Member, said: "The integration of laser radar and vision is not a multiple choice question, but a must answer question. The breakthrough of domestic laser radar in 2024 will reduce the proportion of single vehicle costs from 3% to 0.8%, and the technology maturity curve has crossed the critical point."
The latest report from the Society of Automotive Engineers (SAE) shows that there will be 47 pre installed LiDAR models worldwide by 2024, a six fold increase from 2021. The delivery volume of vehicle mounted LiDAR is also experiencing explosive growth. Taking Hesai Technology and Sagitar Juchuang as examples, Hesai Technology's laser radar delivery volume officially exceeded 100000 units in December 2024, and it is expected that the planned annual production capacity of laser radar will exceed 2 million units by 2025. Sagitar Juchuang's cumulative sales of LiDAR in the first three quarters of 2024 also reached 381900 units, a year-on-year increase of 259.6%. The rapid growth in sales undoubtedly indicates strong demand in the LiDAR market.
The breakthrough path of LiDAR from "luxury" to "standard" is not only a market choice, but also a downward trend in LiDAR prices.
According to the big data research and analysis of Autohome, among the numerous factors that users consider when buying a car, 82% of users indicate that safety is the primary consideration. Meanwhile, the "2023 China New Energy Vehicle Safety Development Report" by Autohome Research Institute shows that over 55% of users agree that equipping with LiDAR is a safer route for autonomous driving technology.
The downward trend of LiDAR prices is mainly due to the reduction of research and development production costs, economies of scale, and localization.
For example, through vertical integration of the industrial chain, the VCSEL laser production line jointly built by Hesai Technology and Sanan Optoelectronics has reduced the cost of emission modules by 65%; The SPAD chip developed by Sagitar Juchuang has reduced the cost of the receiving end to one-third of its competitors. The average unit price of Hesai Technology's products has significantly decreased from 35869 yuan/unit in Q2 2022 to 4023 yuan/unit in Q3 2024, a decrease of nearly 90%. The average price of ADAS LiDAR from Sagitar Juchuang has also decreased from 225000 yuan in 2020 to 26000 yuan in the first quarter of 2024, with a price drop of nearly 90%. The sharp drop in prices has greatly reduced the cost of installing LiDAR, promoting its application in fields such as autonomous driving.
In addition, by 2024, China's annual production capacity of LiDAR will exceed 2 million units, and economies of scale will drive unit prices into the "$300 era". When the unit price of LiDAR exceeds the threshold of $200, its market penetration rate will show exponential growth. In addition, the pan robot market is also opening up a second battlefield. New species such as delivery robots and industrial AGVs are reshaping the LiDAR market landscape.
It is worth mentioning that the Chinese market has become an important growth engine for LiDAR. In 2024, the front-end installation rate of vehicle mounted LiDAR will exceed 8%, achieving a leapfrog development compared to 1.2% in 2022. At the same time, the localization rate of China's LiDAR supply chain has increased from 45% in 2021 to 90% in 2024.
Under the influence of multiple factors, multiple institutions have made positive predictions for the LiDAR industry. Yole predicts that by 2027, the global market size of in car LiDAR will reach 6.2 billion US dollars, with an average annual compound growth rate of 41% and a marginal cost reduction of another 40%.
Trend Force Jibang Consulting predicts that, driven by L3 and more advanced auto drive system systems and logistics transportation, the market output value of lidar is expected to grow from 1.181 billion dollars in 2024 to 5.352 billion dollars in 2029, with a compound annual growth rate of 35%. S&P Global Mobility predicts that the sales of LiDAR will continue to grow in the future, and the demand for LiDAR in the Chinese market will far exceed that of other countries or regions.
When LiDAR shifts from a "luxury configuration" to a "universal safety standard", this revolution driven by cost, technology, and policy is reshaping the global automotive industry landscape. Chinese enterprises, with their advantages in the entire industry chain, rapid iteration capability, and deep understanding of local scenarios, have taken the global high ground in the core sensor field of intelligent driving. With continuous breakthroughs in underlying innovation, LiDAR will not only redefine the perception capabilities of cars, but may also give rise to a trillion dollar scale new ecosystem of smart roads, high-precision maps, and vehicle road cloud integration - in this transformation, China's intelligent manufacturing is transforming from a pursuer to a rule maker.
Source: laserfair
