Autopilot may be a false proposition for a long time to come.

Recently, there have been some hot events about autopilot in the industry. as a practitioner, Lao Li thinks that there should be more neutral voices in the industry to talk about autopilot objectively. The capital market is famous for its red heat on the surface and blue calmness and rationality in its bones. In the face of autopilot, the capital market maintains enough geography.

Many friends have been looking forward to the birth of a new giant in the field of self-driving like the "Ningde era". On the one hand, it can recreate a cash cow, on the other hand, it can also lead the intelligent electric vehicle industry to achieve considerable development and benefits. Lao Li's team also studied it for a long time in the first half of this year, and the answer is no. It is not only difficult for the field of self-driving to have a cash cow like the Ningde era. What's more, there will not be evergreen concept stocks like those in the Ningde era.

Today, Lao Li and you describe the industry chain of self-driving. Why does the secondary market think that the field of self-driving can no longer give birth to a "Ningde era"? Where will autopilot go in the eyes of capital?

These three systems are not the other three systems.

The self-driving industry chain is the longest, widest and most diversified chain in the automobile industry.

Lao Li mentioned in his article "how to stir up Automobile stocks" that one of the most important directions of capital research is the study of industrial chain and value chain, which judge the potential of industrial chain and value chain through supply relationship. today, Lao Li uses this idea to analyze the two chains of self-driving.

Like the three-electric system (battery, motor and electric control), which is the core component of the electrified direction, autopilot can also be divided into three systems, namely, perception system, decision-making system and execution system, no matter what the technical route is, basically can not do without these three systems.

Although the classification is the same, the valuation logic of the two is very different. Electrification is due to the relatively mature technical route, for example, the battery technology is mainly based on ternary lithium battery and lithium iron phosphate battery, which will cover solid-state battery in the future. motor technology is mainly permanent magnet synchronous motor and AC asynchronous motor. The core components of electronic control gradually transition from IGBT to SiC. Under the mature technology system, the relationship between supply and demand has a direct impact on the valuation. Last year and this year, the sharp rise in the lithium industry chain is driven by the relationship between supply and demand.

The logic of autopilot is another logic, and its different technical routes correspond to different industrial chains and leading companies. As the technology is not mature, so the management can not determine the final route, there is no way to determine the progress of mass production, so far, the valuation of self-driving companies is to cross the river by feeling the stones.

Some core components of the three systems are relatively clear, and the perception system can be divided into two parts: vehicle positioning and environmental perception, in which vehicle positioning includes IMU (Inertial Measurement Unit) and GNSS global satellite navigation, and environmental sensing mainly includes camera, ultrasonic radar, millimeter wave radar and lidar.

In the past, people focused on whether lidar is a necessary part for self-driving (Tesla, Inc. thinks it is unnecessary, other car companies think it is necessary). Since last year, almost all institutions think that lidar is necessary, because Tesla, Inc. also began to test lidar in Silicon Valley, high price * high output = big market, lidar is a must.

Each part upstream has a separate supply chain, in the car camera supply chain, the most key component is the CMOS sensor, to put it simply, the camera chip, the domestic listed company Weir shares acquired OV to become the field leader, good performance superimposed semiconductor prosperity, Weir shares skyrocketed, the secondary market industry research is to find more companies like Weir shares.

The decision-making system is N times more complex than the perceptual system. Lao Li's team researchers have made incomplete statistics according to the technical route, including intelligent computing platform, domain controller, high-precision map, information security, V2X, OTA, tool chain, etc., including software and hardware.

The core of the hardware is the autopilot chip, the secondary market has been paying attention to, watching the United States NVIDIA Corp share price soaring, we can only secretly look forward to the domestic alternative concept of the horizon and black sesame listed early.

The core of the software is system software, and friends who have studied American stocks of BlackBerry will find that its share price has soared for a while precisely because an organization has discovered the application scenario of its software system in the field of self-driving, thus driving up valuations.

The execution system looks very simple, but it is actually very difficult. Autopilot has been raising the concept of wire-controlled chassis, the so-called wire-controlled chassis actually includes by-line drive, by-line braking, by-line steering, by-line suspension and so on, leading enterprises may break out in every field, but the problem is that it is not clear who is likely to break out and when. Take on-line steering as an example, Nestor, a listed company in Hong Kong, has risen sharply for a period of time because institutions have excavated on-line steering technology, and the same domestic companies have Bethel and so on.

Although organizations have been mining the companies of these three systems, due to the unclear development and scattered distribution of the industry, people have not been able to draw an industry-recognized autopilot investment map, which is the biggest pain point of autopilot investment.

Gathering is a fire, scattered is full of stars

The disorganized industry chain of autopilot has given researchers a headache, and what makes it even more troublesome for industry researchers is that the value chain of autopilot is also very scattered and difficult to price. These headaches encountered by investors are essentially caused by the immaturity of the industry. Autopilot just seems to be a complete value system, but it is actually complicated.

The market price of an economical electric vehicle battery system is basically more than 60,000 yuan, the market price of the motor system can be controlled within 10,000, and the electronic control system can also be controlled within 10,000, and the value distribution is obvious and concentrated, so it is easy to make a valuation. 60,000 bicycles determine the Ningde era value of 1 trillion, which is why Ningde era has been favored.

The field of self-driving is a different story. No institution in the world should be able to judge the value distribution of the three major autopilot systems. The reason is that the technical route and the pace of mass production are not mature, and the prices of various systems and parts are not mature. The value distribution is changing all the time.

The highest price in the perception system is lidar. The average price of a mechanical rotary lidar in Wesai Technology is about 110000 yuan in 2019 and dropped to about 90,000 in 2020. The prism and rotating mirror lidar of Huawei and Dajiang are thousands to 10,000. In the future, after the large-scale production of MEMS lidar, the cost can be reduced to a few hundred yuan. With a small lidar, the price distribution is so scattered that it is easy to understand why no one can make the value chain.

Decision-making system due to a large number of software and hardware, value evaluation is more difficult, Lao Li can provide several figures for your reference, Mobileye provides extreme krypton 001 visual solution price of about 20,000 yuan, Huawei provides Changan new high-end brand autopilot system price of about 20,000 yuan, XPeng Inc. P7 NGP program cost about 50,000 to 60,000 yuan, different functions, prices vary greatly.

Due to the fierce competition, the cost of the software is also declining. Take high-precision maps as an example. At present, Baidu, Inc. and Gaode's high-precision maps charge for a single year. In the past, Baidu, Inc. 's annual charge for a bicycle was about 500 yuan. Slowly dropped to 200 yuan, or even to 100 yuan. Lao Li personally thinks that there may be no charge in the future.

It is very difficult to implement each on-line control technology in the system, with low production and high price. Lao Li has done a lot of interviews, and no expert can judge a relatively fair price.

Seeing the distribution of the industrial chain and the distribution of the value chain, it is not difficult to understand why it is difficult to give birth to a Ningde era in the field of self-driving. Lao Li himself made a calculation, the value of the motorized bike is at least 80,000 yuan (battery 60,000, motor 10,000, electric control 10,000), and there is little room for decline.

After self-driving is mature, the value of the bike is very difficult to achieve 80,000 yuan, the software part is very difficult to appear oligarchy, and its cost is destined to be very low. Although the cost of the hardware part is high, it is much more competitive than in the field of electrification.

Before Huawei and DJI came in, many researchers had been talking about self-driving. Huawei's entry into the field of self-driving has dealt a big blow to everyone. Huawei's entry into the field of self-driving is a blessing for car companies and users, but a nightmare for capital.

Huawei makes lidar, and its cost and price are lower than those of Wesai Technology and Suiteng Juchuang. Huawei is better at making chips, technology and applications than Horizon and Black Sesame. What is more frightening is that Huawei has also invested in and acquired some upstream supply chain companies. Huawei's MDC product line and ADS product line are all Huawei-related products, so upstream small start-ups are bound to be suppressed. It is even difficult for the companies that may come out to go public, and there is no more opportunity in the secondary market.

Where is the future of autopilot investment?

Some time ago, Lao Li said that the future of autopilot investment is "wait", "wait" is not the goal, the aim is to find market opportunities while waiting.

Lao Li does not know when autopilot will fully break out, but through the development cycle of technology and the company, combined with the evolution time of electrified capital, it has been judged that the intelligent start-up cycle should be 2022-2024. The intelligent leader is autopilot, and once there is a change in the market, Lao Li will update it with you:

The first move in the market is the mass production rhythm of the key link of the self-driving industry chain. The first two sections mentioned that the industrial chain and value chain of autopilot are relatively scattered, in the scattered target, the secondary market can only be weak water 3000, only take a few ladles to drink.

The closest link to listing is actually lidar. In the second half of last year and the first half of this year, many lidar enterprises listed in the United States, and their valuations rose briefly; in the first half of this year, domestic lidar leading Wesai Technology also submitted Science and Technology Innovation Board's prospectus, but failed to list successfully due to financial and other reasons.

Many friends think that enterprises such as Wesai Technology and Suiteng Juchuang will try to land A shares in a variety of ways in the second half of the year. As long as the market boom is good, they are likely to pull out several trading limits in succession. In the upper reaches of the lidar industry chain, there are also some companies that are about to go public, such as Juguang Technology and so on. Although these companies are also the industry leaders in the future, it is difficult to judge the market performance in the short term because of the small overall market. In addition to the most popular lidar link, companies such as vertical vision technology in the perceptual layer also need to follow for a long time.

The decision-making level is the field where giants are most likely to appear in the future, including whole car enterprises, self-driving software enterprises, self-driving hardware enterprises and so on.

For example, companies such as Xiaoma Zhixing and Wenyuan Zhixing all focus on self-driving solutions; companies such as Momenta take autopilot algorithms as the main direction; Geely, SAIC and other enterprises have also applied Momenta algorithms one after another. Recently, capital has been unable to find application scenarios to suppress the valuation of Al. Accordingly, those Al companies that have found landing application scenarios will have higher valuations. Momenta is a typical example.

The second move in the market is the application of autopilot segmentation scenarios. Due to the lack of communication, when users mention autopilot, it is generally understood as passenger car L3 or above application scenarios in a narrow sense, but in fact, from low speed to high speed, from L1 to L5, there are a lot of application scenarios in a broad sense, and these application scenarios are different in safety standards and difficulties in implementation. at present, autopilot in closed scenes is the easiest to achieve, such as the application of mines and docks. The most typical is the wharf autopilot of Yangshan Port, followed by low-level high-speed autopilot, such as commercial vehicle high-speed scene, passenger car high-speed scene, and finally high-level full-speed scene.

Each scene has its own industrial chain, and its technology and cost are slightly different from the corresponding leading companies. for example, the cost of self-driving in a closed scene is more important than the technology, and the cost per bike is basically controlled at 20,000 yuan or even lower. Its self-driving function is also weak, while the high-end self-driving cost of passenger cars is basically more than 60,000 yuan, and the function is more perfect. This not only increases the research scope of researchers, but also puts forward higher requirements for the grasp of investment nodes.

When a signal appears in the market, the popularity of autopilot will gradually rise. the difference is that when the first signal appears, the valuation of the corresponding company is highly speculated, and then there is a mass production divergence in the market. its valuation is likely to return to a relatively normal level. When the second signal appears, the autopilot of the subdivided scene will enter the state of mass production, the valuation of the corresponding industry chain companies will be greatly increased, the market will be highly consistent, and the valuation will not fall.

If these two signals appear in the market at the same time, it is the true output of value, there are value points and mass production applications, which is the most desired state.

Gu Cheng said in "Generation" that the night gave me black eyes, but I used it to find the light. Although high-end full-speed autopilot will be a false proposition for a long time in the future, and although it is difficult to replicate a Ningde era in this field, today we are still looking for a dawn in the bleak and chaotic self-driving industry. It's time for autopilot to take off.