Last weekend, the Chinese team of the US lidar manufacturer Velodyne received a formal notice from the US headquarters, announcing that it would cancel the entire direct sales team and some technical teams in China, lay off more than 20 people in its China office, and only retain a few responsible for channels and after-sales services of major customers Personnel, and restored the product sales model from the direct sales model to the "agent model" when it first entered China. This also seems to mean that Velodyne has actively "abandoned" the Chinese market. Behind this, is the victory of China's lidar manufacturers? Or is the "cold winter" of autonomous driving coming?
Velodyne, the originator of the lidar industry, once occupied 80% of the market
Lidar, which is called Light Detection And Ranging in English, is called LiDAR for short. It is a system that combines three technologies of laser, global positioning system (GPS) and IMU (Inertial Measurement Unit). For obtaining data and generating accurate DEMs (digital elevation models). Because each laser emitted by it has relative position information, combined with the algorithm, it can generate high-precision three-dimensional position information of the surrounding environment of the car in real time, which can accurately and quickly track and identify the target. The human eye can determine the position, size, shape, and even material of surrounding objects.
According to the classification of automatic driving L1-L5, L3 and below belong to assisted driving or low-level automatic driving, and L4 and L5 are considered high-level automatic driving or even unmanned driving. At present, autonomous driving is at a stage below L4, and relying on millimeter-wave radar and cameras can meet the visual requirements of automobiles. However, if you want to develop to the L4 and L5 level of autonomous driving, a lidar capable of generating high-precision 3D point cloud maps of the surrounding environment in real time is more advantageous.
The American lidar giant Velodyne, founded in 1983, is known as the originator of lidar. As early as more than a decade ago, Velodyne's lidar was used in drone cars. In the 2007 DARPA (U.S. Defense Advanced Research Projects Agency) Challenge, six of the seven teams participating in the competition adopted lidar designed by Velodyne. In 2009, Google launched Waymo's self-driving car project. The prototype of its self-driving car uses Velodyne's lidar.
Thanks to decades of accumulation in the field of lidar, Velodyne quickly established a strong advantage in the field of car lidar, once occupying 80% of the market share. With the rise of autonomous driving, Tencent, Baidu, Ford, Mercedes and other manufacturers have launched strategic cooperation with them, of which Ford and Baidu have also invested in Velodyne. Especially in China, where the development of autonomous driving is very hot, Velodyne's lidar products were once the only choice for auto driving technology manufacturers. This is also an important reason why Velodyne previously cancelled the agency model and switched to the direct sales model.
The rapid rise of domestic lidar manufacturers
According to the Boston Consulting Group (BCG), by 2035, the advent of the era of high-tech Internet cars is expected to create about $ 150 billion in new profits for the automotive industry, and this technology is also likely to become a profitable "fragrant tea" ". Some investors estimate that by 2030, autonomous vehicles will account for one-fifth of new car sales worldwide. The optimistic outlook for the future market of autonomous driving has also directly attracted an influx of capital.
BCG data also shows that, although only a few startups may laugh to the end, since 2014, investors have invested $ 70 billion in private investment funds into more than 3,400 companies involved in the "new transportation" concept. These companies Covers many related fields such as autonomous driving (including lidar), ride-hailing services, electric scooters and machine learning technologies.
According to the forecast of foreign research institute Yole, the total number of self-driving cars will reach 44,000 by 2021. At the same time, the lidar market, a key technology for autonomous driving, has also benefited from this, and is expected to grow from $ 1.3 billion in 2018 to $ 6 billion in 2024. Among them, automotive applications account for 70% of the lidar market. Therefore, it has also attracted many capitals into the field of lidar.
With the help of capital, a large number of start-up companies have also emerged in the field of lidar, such as Hesai, Sagitar Juchuang, Divo-invested Livox, Radium God, innovusion and other companies. These startups also hope to use the power of capital to quickly improve their technological strength and market competitiveness, so that they can be in an advantageous position when they can compete with financially powerful auto giants in the future.
With the entry of many domestic lidar startups, after several years of development, not only has the gap between lidar technology and Velodyne been rapidly narrowed (especially, it has an advantage in solid-state lidar technology), but also the cost and local It is also more advantageous in the field of automated services, so it has also won the favor of many autonomous driving technology manufacturers. At present, domestic LiDAR startups with higher financing and valuations are mainly Hesai and Sagitar Juchuang. Hesai's lidar has been adopted by Baidu, Cruise, Nuro, Zoox and other autonomous driving companies, and Sagitar's lidar has also been adopted by domestic and foreign manufacturers such as FAW Group and AutoX. In contrast, Velodyne's products have high prices, unclear channels, and lack of after-sales services. This also makes Velodyne gradually lose its advantage in the highly competitive Chinese market.
For example, Velodyne's 128-wire mechanical lidar is indeed very powerful, but the retail price is as high as 700,000. In contrast, the price-performance ratio of domestic lidar is much higher, and the technology update is more timely. Therefore, we can see that, as one of Velodyne's largest investors, Baidu has also replaced Velodyne's 64-line lidar in this car with Hesai's 40-line lidar from this year. In addition, the trip to Wenyuan, a Chinese self-driving unicorn company with the commercial goal of operating an L4-class self-driving taxi fleet, has also replaced the Velodyne lidar on the car with a Hesai product.
Of course, Velodyne didn't sit still. Earlier, when Velodyne reported the IPO, Velodyne sued Chinese Lidar startups Hesai and Sagitar Jutron in a US court on the grounds of infringing its US patent US 7,969,558.
Earlier, Velodyne used this patent to bring another Lidar Quanergy to court, and the lawsuit has been in progress for more than 4 years. Quanergy filed an invalid application for the patent in November 2017. However, in May this year, the US Patent Trial and Appeal Board ruled to maintain the validity of the patent, Velodyne won the case, but Quanergy sued. At present, the two sides have reached an out-of-court settlement, but the specific terms of the settlement have not been announced.
Cornering overtaking with solid-state lidar
However, Velodyne's US 7,969,558 patent is primarily directed to mechanical rotary lidar products. Although the current mainstream of the lidar market is still mechanical, solid-state lidar, especially solid-state lidar based on MEMS micro-mirrors, is rapidly emerging, and this is also an opportunity to circumvent the Velodyne patent and achieve cornering overtaking.
The mechanical lidar mainly realizes the 360 ° detection of the surrounding environment through the mechanical 360 ° rotation. However, these mechanical parts are relatively complicated, cumbersome, low in production efficiency and high in cost. In addition, these rotating parts will wear out for a long period of time and are prone to failure. In addition, lidar has very high precision requirements (typical lidar applications require angular accuracy of less than 0.5 degrees), and such rotating parts have a great impact on precision. This problem has also stimulated the demand for solid-state lidar.
MEMS solid-state lidar uses a MEMS micro-mirror scanning scheme. Only a few laser transmitters and receivers are needed. Scanning is performed by the MEMS micro-mirror in two directions. Compared with traditional mechanical lidars, MEMS solid-state lidars do not require mechanical rotating parts, can be greatly compressed in structure and size, and their life can be greatly improved. The working principle of MEMS lidar is to use a small number of laser transmitters (16 transmitters are required for traditional 16 lines, and 32 transmitters are required for 32 lines). The MEMS vibration amplitude is superposed to increase the number of scanned lines. MEMS solid-state lidar has higher resolution and longer recognition distance. The data shows that MEMS solid-state lidar can achieve a horizontal resolution of 0.09 degrees and a vertical resolution of 0.2 degrees, and can recognize 10 × 10 cm size 200 meters away Objects. Moreover, it can be arbitrarily pointed within the allowable angle range, and can perform high-density scanning in key areas.
In addition, the silicon-based MEMS micro-mirror can be mass-produced through existing mature semiconductor manufacturing processes, and the quality, consistency, and production capacity are guaranteed. This also makes the solid-state lidar based on MEMS micro-mirror more advantageous in terms of mass production and cost. And for automotive front-mounted components, MEMS solid-state lidar is also easier to meet car-grade standards. At the same time, the MEMS lidar that can be integrated into the interior of the vehicle body can make the whole vehicle more beautiful than the mechanical lidar that needs to be placed on the roof.
In this context, domestic lidar manufacturers have also focused on MEMS solid-state lidar. For example, Hesai and Sagitar Juchuang launched several solid-state lidar products in the past two years. In addition, in September this year, FAW Group signed a cooperation agreement with Sagitar Juchuang to reach a smart solid-state lidar vehicle-level mass production R & D cooperation. It can be said that solid-state lidar has become an important direction for domestic lidar manufacturers to overtake the Velodyne curve.
The "cold winter" of autonomous driving?
It can be said that the rise of domestic lidar manufacturers may be an important factor that prompted Velodyne to "withdraw" from the Chinese market, and another factor that cannot be ignored is that the autonomous driving industry has begun to "fever".
Although autonomous driving is the future trend, at present, autonomous driving is still in the early stages of development, and the technology is not mature enough. Earlier, Waymo CEO John Krafcik said in an interview that although self-driving cars are "really here" and self-driving cars have also appeared in reality, they are far from universal. The main reason is that the autonomous driving technology has not yet reached the highest level L5 that can drive in any weather and under any conditions.
On the other hand, the current policies, laws, regulations, insurance, and supporting infrastructure for autonomous driving are still incomplete, and there is still a long way to go before the real commercial use of autonomous driving is completed.
Another key issue is the issue of trust, that is, the acceptance and trust of human drivers and passengers for autonomous driving. After all, they need to hand over their lives to “autonomous driving”, which requires a long transition the process of. Apple co-founder Steve Wozniak has said that self-driving cars cannot be achieved in the near future, "I don't believe in self-driving cars."
Many analysts believe that it may take decades for self-driving cars to become widespread on the road.
According to the “Autonomous Vehicles Readiness Index” released by KPMG in February this year, from the perspective of the state ’s readiness and openness towards autonomous driving technology, Europe ranks among the top ten countries. 6 seats. But the number one in the Netherlands is only 25.05. China ranks 20th with a total score of 14.41.
Although the future market of autonomous driving is very promising, at this stage, it is still in the "burning money" stage, and it is unlikely that it will continue for a long time. Even in the field of logistics transfer in the closed environment where autonomous driving may land the fastest, there are not many real landings at present, and the demand is relatively limited. In addition, in the field of "RoboTaxi", which is expected to achieve large-scale deployment earlier, more pilots are currently in progress, and progress is slow.
Daimler's new CEO Ola Kallenius has previously stated that the company will "appropriately adjust" its level of spending on autonomous driving technology. Kang Linsong believes that the full deployment of autonomous taxis will take up a lot of capital, and there is uncertainty about the profit potential. He said: "At this stage, it doesn't make sense to be the first company to provide large-scale autonomous taxi services. We will not stop the development of autonomous taxis. We will enter the market at the right time, But we don't think this is work that needs to be completed in the short or medium term. "
On the other hand, the global automobile market has experienced a sharp decline this year, which has also made automakers' investment in the field of autonomous driving also begin to become conservative. According to the forecast of the German Automobile Industry Association (VDA), global car sales in 2019 will fall by 4.1 million to 80.1 million, a year-on-year decrease of 5%. This will be the largest decline in the global auto industry since the financial crisis in 2009. At the same time, VDA predicts that the global automobile sales decline trend will continue in 2020, which is expected to fall by 1% to 78.9 million units.
In addition, in the cold winter of the automobile market, many “new forces” in automobile manufacturing have fallen into a financial crisis. In October this year, there were rumors that four car companies, Cheetah, Zotye, Huatai, and Lifan, will enter bankruptcy proceedings by the end of the year. Although these companies have issued statements to deny the rumors, it is still difficult to hide the downturn.
It can be seen that although the current autonomous driving field is still hot, more of this heat comes from the attention of startups, investment institutions and the outside world to new technologies, but the actual market demand is very limited. For many car companies and autonomous driving technology manufacturers' own testing needs.
Similarly, the current demand for lidar is mainly from the testing needs of autonomous driving technology manufacturers. This also makes the current real market demand for lidar relatively limited. According to a set of figures released by Velodyne in March 2019, sales of Velodyne lidar have exceeded 30,000 units in 12 years since 2007. Other start-up lidar manufacturers are likely to ship fewer. Although from a revenue perspective, Velodyne's revenue has reached 500 million US dollars, but this is all based on the high price of tens of thousands of dollars for a single product, and the current price of lidar is falling sharply, but the market demand The amount has not increased correspondingly.
It should be pointed out that although many large autonomous driving technology manufacturers believe that lidar is the key technology to achieve L4 / L5 automatic driving, many large autonomous driving technology manufacturers are not optimistic about lidar. For example Tesla and Mobileye. As we all know, Mobileye mainly relies on cameras to achieve autonomous driving, and Tesla, the leader in the field of new energy vehicles, mainly relies on cameras + millimeter-wave radar to achieve autonomous driving. At the previous Tesla "Autonomous Driving Day" event, Tesla CEO Elon Musk also said in an amazing statement: "Any autopilot company using lidar is doomed." Another major new energy vehicle manufacturer, Weilai, has also chosen to cooperate with Mobileye this year, and has recently laid off its North American autonomous driving team.
In addition, some autopilot manufacturers or automakers who are optimistic about lidar have also started to develop lidars. For example, Waymo, the global leader in autonomous driving, developed its own lidar and announced that it will sell its lidar Honeycomb to non-RoboTaxi competitors. And the international first-class supplier Bosch is also developing lidar. In October this year, Huawei also announced that it was developing lidar. Obviously, this will undoubtedly be a big blow for third-party lidar manufacturers represented by Velodyne.
Taken together, the autonomous driving and lidar markets are currently in the incubation period, but with the entry of giants and the boost of capital, competition in this field has entered a fierce heat. It is foreseeable that before the era of autonomous driving really comes, players in the field of lidar may be shuffled.