Investment opportunities in automotive electronics brought about by intelligent transformation

In recent years, the two biggest changes in the automotive industry are engine change and steering wheel change, which correspond to the electrification and intelligence of automobiles respectively.

Compared with electrification (new energy), the development of automotive intelligence industry is slower than others, due to the industry attributes and technical characteristics. Although the new energy sector has not yet reached its peak, the sector investment has already made some advance. In comparison, the automotive intelligence sector undoubtedly has a higher investment cost-effectiveness.

First, here is an overview picture:

Of course, it is not just smart driving and domain controller related companies that benefit from the development of automotive intelligence, such as air suspension, AI chips, and wire-controlled chassis. There are more than one company that will grow and develop with the logic of intelligence and domestic substitution. Due to limited personal energy, this article only conducts a simple logical combing of the most direct beneficiaries of smart driving and domain controllers, aiming to record an entry point for studying the intelligent industry chain.

Intelligent Driving Overview

In the past few years, when smart driving is in full swing, many companies will mention mass production of L3 when talking about smart driving. Is L3 smart driving in the strict sense really that easy to achieve? Most investors do not have a very specific concept of the functional level that L3 is to achieve. In fact, this concept is a standard for the degree of intelligence of automobiles proposed by the American Society of Automotive Engineers, the world's most authoritative organization in the automotive industry. The standard is as follows:

The definition of L3 is as follows: the unmanned driving system completes all driving operations, and the driver provides necessary driving intervention according to the system's request. It can be seen that so far, none of the mass-produced models commonly seen in our daily lives meet the L3 definition. If the scope is extended to public road test videos, Huawei's previously announced solution can barely be counted as one, and mass-produced models are all lagging behind and have not achieved true L3 intelligent driving.

The smart driving industry can now be divided into two routes. One is represented by companies such as Bosch, Aptiv, Tesla, and Xiaopeng Motors, which aim to mass produce and gradually iterate from L2 to L3 and L4. The other is represented by Internet companies such as Waymo and Apollo, which skip L3 and directly aim for L4 or even L5. At present, mass production is still dominated by the former, while the latter is more of a small batch trial in a fixed area (such as a park).

The year 2023 is a relatively important time node, because the intelligent driving solutions of major automakers have begun to be mass-produced and launched on the market. It is undoubtedly a year full of challenges and opportunities for all companies in the industry chain.

Before sorting out the core investment sectors of the automotive intelligent industry chain, let’s briefly introduce several core steps for cars to achieve intelligent driving:

1) Calibration: Get the precise position of sensors such as camera, RTK, and lidar relative to the center of the vehicle’s rear axle;

2) Perception: Cameras and lidar detect lane lines, obstacles, traffic lights, signs, etc.

3) Fusion: Information from multiple sensors is integrated and complementary to avoid the defects of a single sensor and improve perception accuracy;

4) Positioning: Obtain the vehicle’s precise location information to facilitate path planning and control;

5) Planning: Based on the perception and positioning output information, the optimal path is planned, such as lane changing, following a vehicle, turning, etc.

6) Control: While controlling the vehicle to travel along the planned path, it also takes into account vehicle stability and comfort.

There are two types of companies in the market that benefit most directly from the development of intelligence: one is to directly output intelligent driving solutions as a supplier. For example, Internet companies represented by Huawei and Baidu, startups represented by Momenta and Pony.ai, and listed companies represented by Desay SV, Thundersoft, Huayang Group, Joyson Electronics, and Jingwei Hirain. Some companies are mainly software-based, while others combine software and hardware. We will not discuss this in detail here, and will only briefly introduce it in the "Domain Controller" section below. The other type is companies that bring incremental parts and products in the process of automotive intelligence upgrades. For example, laser radar companies represented by Hesai Technology and RoboSense, car camera companies represented by Sunny Optical Technology and Lianchuang Electronics, and positioning product companies represented by Huace Navigation and Daoyuan Electronics.

camera

At present, many research reports from institutions estimate that the number of on-board cameras per vehicle will decrease from less than 6 in 2020 to less than 12 in 2025. The following figure shows the market size and growth rate of on-board cameras in the world and China (including forecasts) given by ICVTank:

Among them, by 2025, the compound growth rate of the global automotive camera market will be 16%, and the compound growth rate of the Chinese automotive camera market will be 32.7%.

But the reality is that the new models that are about to be launched and claim to be smart driving will have at least 8-megapixel on-board cameras per vehicle, and generally 12-megapixel cameras per vehicle are the mainstream. Therefore, I personally feel that some organizations have indeed overestimated the strength of the algorithm and underestimated the market size of on-board cameras. The cost of a camera per vehicle is about two to three thousand yuan per vehicle (the specific data is related to the positioning of the vehicle model). There will be a discount in mass production, but the order of magnitude of the cost per vehicle will not change much. At present, mass-produced vehicles with smart driving are in the initial stage, and large-scale mass production has not yet begun. Therefore, this market will have a high degree of certainty in the next few years.

Core A-share targets: Sunny Optical Technology, Lianchuang Electronics, and Novel Optics.

Sunny Optical Technology is the leader in the field of automotive cameras in Hong Kong stocks. Hikvision also has a wide layout, but the proportion of automotive cameras in its main business is too low. For Hikvision, which has a market value of hundreds of billions, the impact of automotive cameras is not obvious. If you really want to choose an investment target in the field of automotive cameras, I personally think Lianchuang Electronics is better. Although Yongxin Optics is also one of the core companies with the most direct benefits from smart driving, its automotive camera ranking is outside the top five domestic companies. In fact, its most direct logic should be laser radar.

LiDAR

LiDAR, needless to say, is definitely one of the most core and obvious beneficiaries of the trend of vehicle intelligence. Just take a look at the domestic car companies that release new smart driving cars and publish press releases about how many LiDARs can be installed, and you will know the status of LiDAR in intelligent driving.

Development trend of LiDAR industry:

With a compound annual growth rate of more than 60%, it is a high-growth stock. Moreover, this field not only benefits from intelligent driving, but also from multiple intelligent fields such as sweeping robots and intelligent distribution.

According to the classification, laser radar can be divided into three types: mechanical, semi-solid and solid. At present, unmanned logistics vehicles and street sweepers, which are fixed-scene products in parks, tend to be mechanical as the main type, with solid and semi-solid as the auxiliary types. The main applications in passenger cars tend to be solid and semi-solid.

Core laser radar companies at home and abroad:

Unfortunately, we did not find the most core laser radar product companies in the A-share market: none of them are listed. As the leading domestic laser radar company, Hesai Technology had previously submitted an application for listing, but withdrew it for some reasons. It should apply for listing again in the future. Sagitar also seems to have signs in this regard, so we can continue to pay attention. In addition, Sagitar's new business has expanded rapidly in recent years, and it is likely to surpass Hesai.

Although there are no companies listed on the A-share market that make lidar, there are still listed companies upstream that directly benefit from it.

Core A-share targets: Focus Optics, Novel Optics, Tengjing Technology, and Everbright Optics.

Focuslight Technology: laser components, Huawei concept;

NovoOptics: Supplier of Hesai, with additional vehicle-mounted camera concept;

Tengjing Technology: optical components, fiber optic devices;

Everbright Huaxin: laser chip.

There are certainly more companies that will benefit from the expansion of LiDAR. Here we only list a few companies with relatively high recognition. If you are interested, you can explore further.

High-precision positioning

High-precision positioning and high-precision maps are not the same concept, this point must be made clear. Moreover, high-precision positioning is only a relative concept. Compared with the positioning results directly output by GPS and Beidouxing, the accuracy of high-precision positioning can be as high as centimeters. As for the location unit that realizes high-precision positioning, some are called combined navigation, and some are called combined inertial navigation or inertial navigation. Although the names are different, the functions are the same: the purpose is to achieve a higher level of positioning than GPS or Beidouxing.