Why did the laser radar, which was disliked by Musk, become a technological "Internet celebrity"?

In 1916, Einstein proposed the theory of stimulated emission of light, and humans began to understand lasers.

In the following decades, with the advancement of science and technology and the development of the times, human understanding of lasers has gradually deepened, and this facility, which was once only used in military and other fields, has begun to appear in front of the public.

On the evening of September 15, the world's first mass-produced smart car equipped with lidar, Xiaopeng P5, was officially launched. This may also mean that high-level intelligent safety has entered the actual implementation stage and has been put on the agenda.

In fact, in addition to being used in autonomous driving of new energy vehicles, there have been many good news in the field of lidar this year.

In August, Japanese automotive lighting giant Koito increased its investment in American LiDAR company Cepton; coincidentally, domestic solid-state LiDAR supplier Tanwei Technology also completed a RMB 100 million Series A financing in September.

Why is this once forgotten facility attracting so many brands?

The rising popularity of LiDAR

After nearly 44 years of technological progress, TH Maiman developed the first laser transmitter in 1960, which marked the beginning of the use of lasers by humans.

A year later, in 1961, China's first laser transmitter, the "Small Ball Illuminating Ruby", was officially launched in Changchun.

In 1969, the United States landed on the moon for the first time and installed a retroreflector device on the lunar surface. By firing lasers from Earth at the device, humans measured the precise distance between the Earth and the Moon. This was also the first time that lasers were used for other purposes.

In the late 1990s, with the development of artificial intelligence, the concept of autonomous driving has also been rapidly improving. Since 2005, there have been more and more examples of lidar being used in the automotive field. Between 2004 and 2007, the DARPA driverless car challenge organized by foreign institutions all used lidar.

Subsequently, Velodyne, which started out as an audio company, transformed itself into a lidar manufacturer. Since inventing 3D real-time lidar, it has launched lidar products with up to 64 lines.

By 2005, more than 80 lidar companies had been founded, many of which were betting on new technologies, stimulating a high degree of technological diversity in the lidar field.

In 2010, Google, which began to develop driverless cars, also chose the lidar provided by Velodyne.

In the following decade, as the prospects of LiDAR became promising and the importance of LiDAR increased, more and more companies began to develop automotive LiDAR. A number of leading LiDAR companies such as Velodyne LiDAR, Luminar, Ouster, and Valeo emerged abroad, and there were also "gold-digging" players in China such as Hesai Technology, Huawei, and DJI.

Although, from the perspective of the application of LiDAR, its application fields are becoming more and more extensive, including military, medical, communication, industry, aviation and other fields, the automotive field is the fastest in terms of maturity and growth rate.

Especially in the context of the popularity of new energy vehicles and smart cars, companies developing autonomous driving have emerged in batches, which has also accelerated the popularity of lidar, including Baidu, Uber, Cruise, WeRide, Pony.ai, etc.

At present, the entire LiDAR market is full of competition. According to statistics from Qingtong Capital, 46 companies entered the market in 2020 alone, including star companies such as Sagitar and Leishen. However, the current market concentration is not high, and start-ups still have the opportunity to stand out.

The "overt conspiracy" of technology giants to enter the market

As a new type of perception hardware, compared with cameras and millimeter-wave radars, lidar can accurately perceive the three-dimensional information of the surrounding environment, with a detection accuracy within centimeters. This greatly improves the device's ability to identify obstacles and surrounding environments.

The advent of the Internet era and changes in lifestyles, especially the increased demand for smart hardware, have changed people's perception of traditional life. Many new concepts such as smart cities and smart logistics have been proposed, increasing the application of lidar.

According to Yole's forecast, the LiDAR market size will reach US$5.7 billion by 2026, with a compound annual growth rate of 21% from 2020 to 2026. Among the market segments, smart cities will have the fastest growth rate of 35%; self-driving taxis and buses will grow at a rate of 33%; and smart logistics will grow at 23%.

In addition, autonomous driving and advanced assisted driving capabilities have great requirements for the ability to predict and identify the surrounding environment and obstacles ahead. Compared with the current mainstream detection equipment, lidar generally has a longer effective detection distance, which just meets the needs of unmanned driving and intelligent driving.

Especially due to macro-environmental factors, people have been more inclined to private travel in recent years. The reduction in traffic congestion will also accelerate the trend from public transportation to private driving, and pay more attention to the development of lidar, which is beneficial to long-term development.

At present, the automotive LiDAR market is growing at a rate of 111%. By 2026, the market size will exceed US$2.3 billion, accounting for nearly 49% of the total market size. Among them, LiDAR used for ADAS will account for 41%. LiDAR has become a standard feature of mass-produced cars, and the competition with pure vision has been won.

On the other hand, new infrastructure has become an absolute hot spot for domestic investment, entrepreneurship and society in recent years. Smart transportation and vehicle-road collaboration, as an indispensable part, have attracted more attention. At this time, various sensors are needed, such as lidar and cameras to provide environmental information.

As a smart transportation terminal sensor, the performance of lidar determines that it has an indispensable position similar to L4 autonomous driving, and has huge application prospects in vehicle-road collaboration.

Finally, the market has become more rational, and technical solutions have become the focus of attention. For example, IDG Capital's investment in Benewake Photonics and Lightspeed China's investment in Hesai Technology are mainly due to their focus on the lidar technical solutions behind these companies. The continuous maturity of technology has become the key to the influx of companies.

At present, solid-state LiDAR is the industry-recognized development direction. After continuous optimization, the size of the improved solid-state LiDAR has been significantly reduced compared to the original equipment form. Even the size of XenomatiX's product has been reduced by 3 times, but the energy consumed is only half of the original, and the function and effect have increased by 3 times.

More importantly, the price of early mechanical LiDARs was hundreds of thousands or even millions. Now in the solid-state industrial chain system, the quality and price of LiDARs are beginning to reach a reasonable range, which is one of the reasons why major manufacturers love LiDARs.

The tide surges on the waves

In fact, although the industry players are a bit crowded, the industry concentration is not high at present.

Angus Pacala, founder and CEO of Ouster, said, "In the next five years, only three to five lidar companies will survive." This statement seems a bit alarmist, but it also reflects from the side that the trend of industry integration is still continuing.

According to the latest data released by Yole, Valeo, which currently has the highest market share, does not exceed 30%, and among other industry players, only RoboSense ranks second with 10%, while the rest of the players have single-digit market shares, which shows that the market structure is still relatively fragmented and many players still have opportunities.

The implementation of LiDAR, especially automotive LiDAR, is of extraordinary significance. For the domestic LiDAR industry, entering the mass production stage marks the relative maturity of LiDAR technology.

One very obvious sign is that the reason why Musk was not optimistic about LiDAR two years ago was because the price was relatively high at that time, which would lead to an increase in the cost of the entire vehicle, but with the current trend, the price of LiDAR will become cheaper and cheaper.

At present, the average selling price of lidar is around 5,200 yuan. As the technology gradually matures and costs decrease, it is expected that the average selling price of lidar will drop to 1,042 yuan by 2026, which will also greatly improve the overall application efficiency of lidar.

However, looking at the entire field, the large-scale implementation and full-scenario application of domestic LiDAR is still subject to some factors and problems. From an objective point of view, dealing with adverse conditions is a difficulty that LiDAR must face.

For example, in Uber's first fatal self-driving car accident in 2018, the cause was that the self-driving car's sensors had detected the pedestrian crossing the road, but the self-driving software did not take evasive measures at the time.

Ensuring that LiDAR has its proper place in the vehicle is key to determining its fair and reasonable handling of application scenarios, thereby helping the autonomous driving system make faster decisions and achieve or ensure safe vehicle operation.

On the other hand, just like cameras, the basic hardware was the main difference initially, but as the industry develops, differentiation in software and applications becomes crucial. How to achieve differentiation in software and applications may be a key link.

Taking the automotive field as an example, this type of field requires LiDAR to have depth sensing, and at the same time it can also realize the intelligent fusion of many sensors in the car, realize differentiation in software or applications, and be able to accomplish more things with the same number or fewer sensors. It can detect the entire scene in an "instant", provide different fields of view, different ranges, different resolutions and frame rates, and be more flexible.