The automobile intelligent terminal will become the Free nerve ending of the intelligent era. The automobile chip is the core of helping the automobile enter the intelligent era. Big computing power and accurate perception are the basic requirements of automobile intelligence, which will drive the rapid development of semiconductor devices from big computing power core chips to various sensors and radars. Compared to the early relatively simple assisted driving and the currently only ideal L5 level autonomous driving, various intelligent driving and ADAS systems have gradually become widely popular on various vehicle models. It is predicted that the penetration rate of L1/L2 level autonomous driving will reach 50% in 2025, and now it is only 35%. However, the penetration rate of L3 level autonomous driving will rapidly increase from 5% in 2020 to 45% in 2035, and the penetration rate of L4 and above level autonomous driving will also increase from less than 1% now to around 22% in 2035. QY Research predicts that the global intelligent driving related chip market will have a compound annual growth rate (CAGR) of over 20% from 2021 to 2028. The increase in demand for electric vehicles and the popularization of autonomous driving technology are the main factors driving its development. In addition, the growing demand for Advanced Driver Assistance Systems (ADAS) is expected to drive market growth.

Intelligent driving changes the semiconductor landscape of automobiles

By 2025, the penetration rate of high-level autonomous driving will reach 65.5%, and the market demand for intelligent driving chips will reach 14 million sets per year. So which specific categories of chips will benefit from the rapid popularization of autonomous driving technology?

MCU is the most widely used digital chip in automobiles. Traditional cars require an average of 70 MCUs per car, while cars equipped with intelligent driving functions may require more than 300. The application fields include ADAS, body, chassis and safety, information entertainment, powertrain, and so on, almost everywhere. Therefore, the market size of automotive MCU is expected to reach nearly 12 billion US dollars by 2025, corresponding to a compound average growth rate of 14.1% from 2021 to 2025, which is significantly higher than the overall MCU market growth rate of 8% in the next three years.

The trend of automobile intelligence has brought several orders of magnitude of improvement needs to the intelligent architecture and algorithm computing power of automobiles, driving the rapid shift of automobile chips towards SoC chips with stronger computing power. The biggest driving force comes from the rapid popularization of intelligent cockpit technology. Due to the continuous increase in demand for "one core multiple screens", intelligent cockpit chips need to support both information entertainment functions and various dashboard display functions, And we should gradually develop towards miniaturization, integration, and high-performance. As the most intuitive unit for improving driving and riding experience, the intelligent cockpit has become another core competitive advantage of new energy vehicles compared to traditional fuel vehicles. According to IHS statistics, the penetration rate of intelligent cockpit new cars in the global market and the Chinese market is increasing year by year, and it is expected to increase to 59.4% and 75.9% respectively by 2025.

Autonomous driving chips become the most dazzling stars

Autonomous driving chips are the fastest-growing field among smart car chips. Due to the urgent demand from consumers for L3 and above level autonomous driving functions, and some countries having started testing and applying L3/L4 level autonomous driving functions in closed, semi closed, and low-speed scenarios, high-performance autonomous driving chips are bound to usher in a broader market space than before. The current computing power required for cars can be comprehensively calculated based on the amount of data collected by sensors in the car. For L2 level autonomous driving, the computing power requires approximately 10 TOPS. The data shows that L3 level autonomous driving requires 80-100 TOPS of computing power, L4 level autonomous driving requires 600 TOPS or above, and L5 level autonomous driving requires over 2000 TOPS. It is expected that autonomous vehicle will start a round of explosive growth around 2025, and the compound growth rate of the autonomous driving chip market with great computing power between 2023-2030 will not be less than 30%, and gradually become the most economically valuable chip product in vehicle semiconductors. From the overall development trend of autonomous driving, due to the need for a large amount of redundant sensor information to be processed, autonomous driving has extremely high requirements for terminal computing power. At the same time, it puts forward higher requirements for real-time, energy efficiency, and reliability of terminal computing. Overall, dedicated chips will become the mainstream solution for future autonomous driving processor chips.

Intelligent driving cannot do without the application of massive sensors

The core of intelligent driving is the collection and processing of various types of information both inside and outside the car. Beizhesi Consulting predicts that by 2027, the global automotive sensor market will reach 324.376 billion yuan with a CAGR of 13.97%. Especially, various sensor systems that can better serve intelligent driving will become the main force leading the rapid development of the automotive sensor market.

Due to the increasing demand for accurate detection, recognition, transmission, and analysis of environmental information around vehicle bodies, the importance of LiDAR has gradually become prominent. As one of the important configurations of mid to high-end intelligent driving vehicles, LiDAR, according to Yole data, the global shipment volume of LiDAR in the ADAS market will increase from 200000 units in 2020 to 3.4 million units in 2025, with a CAGR of 76.2%; It is expected that the shipment volume will reach 26.6 million by 2032.

Vehicle mounted CIS is the core sensor of ADAS, which can make up for the defects of radar in Object detection, and is also the sensor closest to human vision. It is widely used in the automotive field. CIS has been gradually extended from its early use in driving records, reverse imaging, parking and surround view scenarios to intelligent cockpit behavior recognition and ADAS assisted driving, and its application potential has begun to highlight. Due to the addition of autonomous driving function, the demand for car cameras will rapidly increase, and the corresponding demand for CIS will also significantly increase. If L3 or higher level autonomous driving can be achieved in 2023, the number of single car cameras is expected to increase to around 11 to 16, with a maximum demand of around 1.6 billion cameras and a market size of 8 billion to 10 billion US dollars.

Undoubtedly, the rapid application of intelligent driving has become an irreversible development trend in the automotive industry, which will profoundly change the supply pattern and future development direction of traditional automotive semiconductors. So how to seize the latest trends in the development of the automotive electronics market and grasp the future opportunities of automotive semiconductor design and application?