New energy and intelligent vehicles
After the new energy vehicle market broke out in 2021, it will continue to maintain a high growth trend in the first month of 2022. According to data from the Passenger Federation, the wholesale sales of new energy passenger vehicles in January reached 412,000 units, a year-on-year increase of 141.4%; the penetration rate was as high as 19.0%, an increase of 10.6 percentage points compared to the same period in 2021.
Main documents of this article
Reference link
https://mp.weixin.qq.com/s/0OtzyP252sa9koObULM4Xg
https://baike.baidu.com/item/%E8%87%AA%E5%8A%A8%E9%A9%BE %E9%A9%B6%E6%B1%BD%E8%BD%A6/4881925?fr=aladdin
https://baijiahao.baidu.com/s?id=1725983099368881275&wfr=spider&for=pc
Leading OEMs in active transformation include Geely, BYD, Great Wall, SAIC, etc.;
leading Emma Technology, which has benefited from the replacement of the new national standard for two-wheelers and internal improvement of management; the
hybrid industry chain brings new increments of parts and components. In terms of engine control, Lingdian Electronic Control is recommended. Companies in the three-in-one track include Juyi Technology, Wolong Electric Drive, Jingjin Electric, and Xinrui Technology in the field of Xiaosanhe;
the main participating companies in the thermal management field are Sanhua Intelligent Control and Toppu. Group etc.
In terms of electric drive and electric control, OEMs, traditional Tier 1 suppliers and third-party suppliers compete. In China, relying on the layout of independent vehicle brands and the accumulation of third-party manufacturers, BYD, Weiran Power (NIO) and Honeycomb Power have emerged. (Great Wall), Wei Rui Electric (Geely) and other vehicle brands, as well as third-party manufacturers such as Founder Electric Drive, Inovance Technology, Wolong Electric Drive, and Shanghai Electric Drive, while traditional Tier 1 suppliers are mainly Bosch, Continental, Valeo, Siemens and other overseas giants.
With the rise of independent vehicle brands and domestic suppliers of "three electric systems", it will naturally drive the domestic substitution of electronic components for upstream vehicles.
There are many types and quantities of auto parts, and they involve different industries and fields. There is a big gap in technical standards and production methods. At present, the relevant policies of the state on the auto zero manufacturing industry are mainly distributed in the related policies of the auto industry.
The release of a series of policies for the automobile industry has also put forward higher requirements for the parts industry.
Self-driving automobile (Autonomous vehicles; Self-driving automobile), also known as driverless car, computer-driven car, or wheeled mobile robot, is a kind of intelligent car that realizes unmanned driving through a computer system. In the 20th century, it has a history of several decades, and the beginning of the 21st century shows a trend of close to practical use.
Self-driving cars rely on artificial intelligence, visual computing, radar, surveillance devices, and global positioning systems to work together to allow computers to operate motor vehicles autonomously and safely without any human intervention.
Autopilot technology includes video cameras, radar sensors, and laser rangefinders to understand the surrounding traffic and navigate the road ahead through a detailed map (from a human-driven car). This all happens through Google's data centers, which process the vast amount of information the car collects about the surrounding terrain. In this regard, self-driving cars are the equivalent of remote-controlled cars or smart cars in Google's data centers. One of the applications of Internet of Things technology in automotive autonomous driving technology.
Volvo distinguishes four stages of autonomous driving according to the level of automation: driver assistance, partial automation, high automation, and full automation.
1. Driving Assistance System (DAS): The purpose is to provide assistance to the driver, including providing important or useful driving-related information, as well as clear and concise warnings when the situation begins to become critical. Such as the "Lane Departure Warning" (LDW) system.
2. Partially automated systems: systems that can automatically intervene when the driver receives a warning but fails to take appropriate action in a timely manner, such as the "Automatic Emergency Braking" (AEB) system and the "Emergency Lane Assist" (ELA) system.
3. Highly automated system: A system that can replace the driver to control the vehicle for a long or short period of time, but still requires the driver to monitor the driving activities.
4. Fully automated system: A system that can unmanned a vehicle and allow all occupants in the vehicle to engage in other activities without monitoring. This level of automation allows for computer work, rest and sleep, and other recreational activities.
In-vehicle chips, who is in the first echelon in China?
At present, the sales of domestic chip companies such as Horizon, Huawei, and Black Sesame are mainly based on Chinese self-owned brand passenger cars. Therefore, the market size of Chinese self-owned brand passenger cars is estimated to analyze the development space of these companies. 2025 is a watershed, the functional penetration rate of L1 and L2 will increase to the top in 20 to 25 years, and then the penetration rate above L2 will continue to grow rapidly from 25 years, while the penetration rate of L1 and L2 will gradually decline.
Horizon, Black Sesame, Huawei, Mobileye, Nvidia, the new decade, the two major stages of the market pattern. The content of this article is as follows:
With the continuous improvement of the penetration rate and level of ADAS and DMS functions, the number of sensors and computing power requirements also increase, which directly stimulates the volume and price of automotive AI chips. Facing the vast domestic passenger car market, the participants in the future high-level chip track mainly include domestic Horizon, Black Sesame, Huawei, and foreign Mobileye, NVIDIA and Tesla.
ASIL (Automotive Safety Integration Level), the automotive safety integrity level, describes the probability that the system can achieve the set safety goals. In China, suppliers' components generally need to meet the ASIL B standard in order to be allowed by traditional large car companies as the core components of ADAS and other assisted driving modules. Some more aggressive car companies may accept chips that do not meet ASIL B requirements, but such a solution also needs to increase the redundant design of other chips and components to ensure the functional safety of the entire system.
1. Horizon
Horizon provides intelligent driving solutions through chips based on "algorithm + chip + tool chain" as the basic technology platform, enabling the intelligent Internet of Things. As a secondary supplier, Horizon is responsible for the construction and improvement of the basic technology platform. The product line includes the Journey 2 launched in 2019, the Journey 3 launched in 2020, and the planned Journey 5 launch. It covers from L2 assisted driving, human-computer interaction of intelligent cockpit, to assisted driving and autonomous parking close to L2+. In addition, Journey 5 benchmarks Tesla's FSD, which is aimed at high-level automatic driving and is based on the central electronic and electrical architecture, that is, the chip of the central on-board computer architecture.
Horizon's business model positions itself as a Tier 2 supplier. Customers include Tier 1 suppliers, OEMs, and travel service providers. The solutions provided include reference designs for chips and hardware, as well as tool chains and algorithms. Horizon is a company rooted in China. Its core R&D team is deeply involved in the Chinese market. It believes that having localized service capabilities will help improve Horizon's competitive advantage over foreign competitors.
2. Black Sesame
Black Sesame provides perception system solutions through a combination of neural network visual perception algorithms, vehicle-grade ADAS/autonomous driving chips, supporting underlying implementation systems and reference designs. The company targets Tier 2 and cooperates with automakers and Tier 1 suppliers such as Bosch, SAIC, FAW and General Motors. At present, the black sesame series chip products include Huashan No. 1 A500, Huashan No. 2 A1000 and A1000L. The black sesame Huashan No. 1 A500 chip has begun mass production, and no mass-produced models have been seen so far. In addition, the company expects that domestic models equipped with the Huashan No. 2 A1000 chip will be officially mass-produced by the end of 2021.
Based on the Huashan No. 2 A1000 chip, Black Sesame provides four intelligent driving solutions, which are used to support different levels of assisted driving and automatic driving functions of ADAS, L2+, L3 and L3/L4. In addition, Black Sesame can provide different customized solutions required by customers.
At present, the A1000 of Black Sesame’s chip products can be compared with NVIDIA chips in terms of single-chip computing power, but the highest computing power of 70 TOPS is still far from NVIDIA Orin’s highest single-chip computing power of 200 TOPS. There is also a gap between NVIDIA and NVIDIA in terms of computing platform solutions. The solution for L3/L4 high-level tracks given by Black Sesame uses 4 A1000 chips and has a computing power of 280 TOPS. This is the solution with the highest computing power among all computing platforms at present, but it is lower than the 320 TOPS computing power of the Pegasus platform released by NVIDIA in 2017. In addition, NVIDIA also has the newly released DRIVEAGX Orin autonomous driving platform at the end of 2019. The highest computing power can be Up to 2000 TOPS.
3. Huawei
Huawei relies on its ICT advantages to deeply cultivate the three-dimensional layout of "end (vehicle intelligence and networking equipment), management (vehicle networking infrastructure), and cloud (vehicle networking platform)", positioning Tier 1, and benchmarking against Bosch. And reach alliances with OEMs, parts manufacturers, and software companies, and rely on the core "core" advantage to target the integrator of the domestic smart car industry.
On May 9, 2020, Huawei joined forces with FAW Group, Changan Automobile, Dongfeng Group, SAIC Group, GAC Group, BAIC Group, BYD, Great Wall Motor, Chery Holdings, Jianghuai Automobile, Yutong, Celis, Nanjing Iveco, T3 Travel, etc. The first batch of 18 car companies), officially established the "5G car ecosystem". Huawei has the ability to integrate the resources of various giants, deploy the interconnected equipment required in the research and development process, and speed up the research and development of driverless vehicles. In the future, there may be opportunities to overtake the horizon in the corners.
Judging from the information disclosed so far, Huawei will mainly focus on software/hardware fields such as chips, algorithms, Huawei Cloud, V2X, and operating systems, in addition to other technologies such as car shells, chassis, wheels, and seats. However, if TSMC's foundry is affected by Sino-US frictions, Huawei's transformation of car manufacturing can also highlight its technological advantages.
4. Mobileye
Mobileye is a leader in camera-based image recognition technology, with strong technical barriers, and its market penetration rate in ADAS to L2+ solutions exceeds 70%. Represented by the EyeQ series chips, the cumulative global shipments have exceeded 50 million pieces, basically leading the development direction of the global ADAS industry. In 2019, it held 45 cooperation projects with 26 car companies around the world, and won more than 16 million orders for 22 new models. Mobileye enjoys the benefits of making the cake bigger in the popularity of intelligent driving around the world, and was acquired by Intel for $15 billion in 2017.
Mobileye's technology accumulation is mainly based on machine vision and neural networks, using cameras with sensors and customized neural network system chips to perform object detection tasks on a single hardware platform. This solution is inspired by the visual driving state of human drivers. The system captures road conditions through on-board cameras and adapts and adjusts real-time road conditions at any time.
The domestic Horizon, Black Sesame, and Huawei are growing rapidly. Although their products are currently climbing, they have already carried out a number of cooperation with domestic leading car companies. In addition, these companies are deeply cultivating the domestic market, which is more down-to-earth and has local advantages. Mobileye's black box solution may limit the subsequent functional integration of car companies.
Horizon also has the advantage of data collection in China. As a foreign company, Mobileye is subject to many restrictions in China. It needs to cooperate with domestic companies (such as SAIC and Ziguang), and cannot collect high-precision map data alone. In addition, compared to Mobileye, Horizon Cars are deeply involved in the domestic market and have the advantage of localized services. Although Mobileye's product line is more mature than Horizon at present, Horizon will become Mobileye's biggest threat in the domestic market in the future.
5. NVIDIA
NVIDIA has made little contribution to autonomous driving-related revenue in the past two years. The reason is that the company has been "holding high" and hopes to directly solve the computing power of L4 and above. Although the hardware has been repeatedly iterated, it has relatively high R&D requirements for partners. Strict, coupled with the fact that the industry faced the headwinds that have not developed smoothly in the past two years, the timing of the increase in value has been postponed. However, in recent years, the company is also focusing on L2 to lower the threshold for car companies to adapt, hoping to cut in from the top and expand cooperation with Toyota, Mercedes-Benz, and Continental. The automotive industry is at a critical point in electrification and intelligent transformation, but autonomous driving is one of the biggest tracks for the long-term development of AI.
Nvidia has become a "pioneer" in terms of hardware, computing power and R&D rhythm. However, at present, it only cooperates deeply with some new car-making forces. The time for mainstream car companies to increase their volume starts in 2021. The company still needs to solve the problem of in-vehicle computer Balance bottleneck between computing power and power consumption.
Hardware: NVIDIA's main products include Xavier and Orin chips, as well as autonomous driving platforms DRIVE PX Pegasus and DRIVEAGX Orin based on the two, respectively. For L3 and above tracks, it has an absolute leading edge in computing power. The autonomous driving platform DRIVE AGX Orin, released at the end of 2019, has a single-chip (Orin) computing power of up to 200 TOPS, and it has landed on the level of autonomous driving, from the L2-level single-camera Orin S to the L5-level 2 Orin + 2 GPUs Chip, the computing power is increased from 36 TOPS to 2000 TOPS, and the power consumption is increased from 15W to 750W.
Software: NVIDIA's complete software computing platform, NVIDIA DRIVE, can help customers build customized applications. NVIDIA DRIVE IX, an open platform for intelligent cockpits, provides customers with comprehensive capabilities, including driver monitoring, natural language processing, and in-vehicle visualization.
NVIDIA's complete software computing platform NVIDIA DRIVE provides a high degree of openness to meet the R&D needs of car companies with strong R&D capabilities. Including the use of active learning, federated learning and transfer learning to train deep neural networks, using NVIDIA as a platform to connect multiple data sets distributed in different countries, creating simulated training scenarios through the Constellation simulation system, and creating an in-vehicle driver monitoring system through DRIVE IX , through DRIVE RC to achieve the same platform from passenger car companies, commercial vehicle companies, Tier One to startups.
NVIDIA is positioned at L3 and above levels of autonomous driving, but the gap from ADAS to L3 level products does not meet the current main needs of domestic car companies. At present, the cooperation between NVIDIA and domestic car companies is mainly in the stage of Prototyping, laying the foundation for the future high-level track. Nvidia said that after 2023, automotive chip products will also penetrate down to levels below L3, enriching the company's product line.
6. The market structure from 2020 to 2030
is expected to be dominated by the ADAS-level automotive chip market from 2020 to 2025. The penetration rate of ADAS functions continues to increase to saturation. The main players include domestic automotive chip manufacturers Horizon, Black Sesame, Huawei, And Mobileye abroad.
Mobileye's products are more mature, with a global penetration rate of 70%, but the black box solution is relatively limited, and the localization service capability is weak; Horizon has a first-mover advantage among domestic players.
It is expected that from 2026 to 2030, the competition will be dominated by L3 and above high-level competition. The functional penetration rate of L1 and L2 will be replaced by L3+ and gradually decline. Major players at home and abroad include Horizon, Huawei, Black Sesame, Nvidia and Mobileye. Among them, we are optimistic about the strength of NVIDIA in the high-level track, but Huawei is positioned as a Tier 1 supplier and builds a "5G automotive ecosystem", which is aimed at the high-level autonomous driving market and cannot be ignored.
Reference link
https://mp.weixin.qq.com/s/0OtzyP252sa9koObULM4Xg
https://baike.baidu.com/item/%E8%87%AA%E5%8A%A8%E9%A9%BE%E9% A9%B6%E6%B1%BD%E8%BD%A6/4881925?fr=aladdin
https://baijiahao.baidu.com/s?id=1725983099368881275&wfr=spider&for=pc