tech

Car companies align intelligent driving capabilities, chip companies and server

In the first half of 2024, despite fierce competition, China's new energy market continued to show growth. According to statistics from the China Passenger Car Association, the retail sales of domestic new energy passenger cars in the first half of the year broke through the 4 million mark, reaching 4.111 million units, a year-on-year increase of 33.1%. Among them, Geely Automobile, Seres Automobile, and Chery Automobile all achieved a year-on-year growth of over 100%.

In addition to the increase in sales, another hot topic in the automotive industry is intelligent driving. The emergence of "Luobo Kuaibao" has brought autonomous driving to the masses. Looking back at the timeline, in November 2023, the Ministry of Industry and Information Technology, the Ministry of Public Security, the Ministry of Housing and Urban-Rural Development, and the Ministry of Transport jointly issued the "Notice on Carrying Out the Pilot of Access and Road Traffic for Intelligent Connected Vehicles," officially launching the pilot for the access and road traffic of intelligent connected vehicles. At that time, some analyses pointed out that this "Notice" was mainly aimed at autonomous driving vehicles above the national standard level 3 (L3), as the document clearly stipulated the relevant requirements for the access of intelligent connected vehicles, the main body of use, road traffic, and the suspension and exit of the pilot.

Advertisement

Although the L3 level is no longer a new topic, this notice seems to have more decisive significance. One of the key factors is that the policy scope is more universal, and cities above the prefecture level (including districts under the jurisdiction of municipalities directly under the Central Government) that meet the requirements can apply for the pilot, marking that the road test of autonomous driving will gradually be carried out nationwide. This is an important background for the implementation of L3, after all, a technology only has the next step when it is used.

The distance is poetry and L3, but the reality is L2.

01

It's 2024, and we are still at the L2 stage.

In the fiercely competitive new energy vehicle market, the domestic market has already "fully L2".Top 10 Domestic New Energy Passenger Car Companies in the First Half of 2024 with L2 Level Intelligent Driving Solutions

In 2023, the penetration rate of L2 level autonomous driving in domestic passenger cars reached 47.3%, and this figure exceeded 50% in May of this year. Currently, since L3 cannot officially hit the road, most intelligent driving companies can achieve the L2 level. Although there are many products on the market such as L2+, L2.5, or L2 advanced. But L2+ is not an official SAE classification, and each manufacturer has different definitions and functions for it. From L0 with no automation to L5 full autonomous driving, L2+ can only represent an attempt to transition from partial autonomous driving to a higher level of autonomous driving, providing enhanced functions beyond the standard L2 level, but the specific implementation depends on the manufacturer.

These Plus versions do not have a fixed standard. No matter how many "plus" signs, they are only infinitely close to L3. These plus signs include traffic congestion assistance, automatic lane changing, navigation assistance driving, and automatic parking, etc., but as long as it is an L2 level intelligent driving, the safety responsibility will fall on the driver.

With policy promotion, the shoe has finally fallen, and L3 has arrived in 2024. Looking back in a few years, 2024 may be the first year of commercialization of L3, and many companies have already obtained L3 test licenses. Mercedes-Benz, BMW, Chang'an Automobile, Ji Hu Automobile, Shenlan Automobile, Avita, and other domestic and foreign car brands have all successfully obtained L3 level test licenses.

Facing the foreseeable trend, the next question in front of us is "What changes will L3 bring?"

02

L3 Intelligent Driving, Good for Hardware and Software Markets

With the development of intelligent driving, the related software market will usher in growth, and urban NOA is one of them. NOA (Navigate on Autopilot) is an automatic assisted navigation driving. Now, all car companies are competing for urban NOA. NOA allows users to achieve point-to-point navigation assistance driving within a specific road range. The vehicle can achieve traffic lights, lane changes, and giving way without being taken over by someone, and finally reach the destination. In September 2022, Xiaopeng took the lead in pushing the urban NOA function in Guangzhou. Currently, more than 20 car brands including "Nio, Xiaopeng, Li Auto," Zhiji, Wenjie, and Jikr have all launched urban NOA high-level intelligent driving. As the name suggests, whoever "unlocks" more cities first can expand the market bigger. The Taibo Research Institute report shows that urban NOA is expected to usher in a turning point of mass production growth in 2026, and the market size is expected to break 100 billion yuan by 2027, with a compound growth rate of 37% from 2023 to 2027.

The update of software will inevitably be transmitted to the hardware level, and on the one hand, the hardware puts forward new requirements for the related chip products on the car; on the other hand, it will also increase the demand for server products in the automotive field.From a technical standpoint, there are several key differences between L3 and L2. Once the L3 system is activated (before the driver takes over), the L3 system assumes the responsibility for safety, hence the safety design requirements for L3 are much higher than those for L2. The design requirements for L3 are significantly higher than for L2, necessitating the consideration of more safety factors, including Safety of the Intended Functionality (SOTIF) and Functional Safety (ISO26262). The functional safety level of L3 systems typically needs to reach ASIL-D, which is the highest safety level in the automotive industry, while L2 systems may not need to reach this level.

Based on these requirements, to meet the safety and performance requirements of L3, more advanced hardware configurations may be needed, such as higher-performance sensors and processors, which also pose higher demands on intelligent driving chips. Therefore, the advent of L3 will inevitably bring new challenges and opportunities to the chip industry.

In the field of intelligent driving chips, domestic products from Horizon Robotics can already cover scenarios from L2 to L4. Overseas, companies like NVIDIA, Mobileye, and Tesla also have relevant products. However, L3 will not only affect intelligent driving chips, but multiple policies are already discussing automotive chip standards under new circumstances. In these policies, it can be seen that various automotive chip products will welcome new standards and requirements.

In July 2023, the "National Vehicle Network Industry Standard System Construction Guide (Intelligent Connected Vehicles)" stated that automotive chip standards mainly focus on the performance requirements and test methods of key chips for intelligent connected vehicles, mainly covering safety chips, computing chips, and other automotive chip-related standards that reflect the characteristics of intelligent connected vehicle applications. The guide also proposed that by 2025, a system will be formed to support the intelligent connected vehicle standard system for general functions of combined driving assistance and autonomous driving, covering standards for combined driving assistance, key systems of autonomous driving, connected basic functions and operating systems, high-performance computing chips, and data applications.

In June 2024, the Ministry of Industry and Information Technology mentioned in the key points of automotive standardization in 2024 that the development of emerging fields in the automotive industry should strengthen the supply of automotive chip standards. This includes accelerating the development of standards for automotive chip environment and reliability, electric vehicle chip environment and reliability, automotive chip information security, providing basic technical support for automotive chips. Promote the formulation of standards for intelligent driving computing chips, automotive ETC chips, infrared thermal imaging chips, cellular communication chips, security chips, power drive chips for electric vehicles, and analog front-end chips for power battery management systems in electric vehicles, clarifying the technical requirements and test methods for various types of chips.

The automotive cloud market is welcoming favorable conditions.

The needs for intelligent driving, such as vehicle networking and data analysis, will generate more data, and the automotive cloud market has been growing significantly since 2023. IDC stated that in the second half of 2023, the overall scale of China's automotive cloud market had reached 5.35 billion yuan, a year-on-year increase of 35.2%. Among them, the infrastructure market scale was 3.16 billion yuan, a year-on-year increase of 43.8%; the solution market scale was 2.19 billion yuan, a year-on-year increase of 24.5%.

With the improvement of intelligent driving levels, the consumption of cloud resources by car companies is greater.

Previously, He Xiaopeng, chairman of Xiaopeng Motors, said: "Recently, I have talked with several leaders of L4 about end-to-end, and they all believe that end-to-end is the best route for L2 driving assistance or L3 autonomous driving, but it is definitely not the preferred choice for L4. I really believe that end-to-end and large models will eventually reach L4 autonomous driving." In order to provide consumers with a better experience, car companies have started to apply large models in intelligent driving experiences, which has further increased the consumption of cloud resources.

Taibo Research Institute predicts that the scale of China's autonomous driving data closed-loop market in 2023 is about 155.4 billion yuan, and it will break through 500 billion yuan in 2030, with a compound annual growth rate of 17.5% from 2023 to 2030. The growth of data volume will also bring about the growth of the automotive cloud market.Has L3 Taken Away Jobs from DiDi Drivers?

It appears that L3 has brought prosperity to various industries, but there are concerns about whether autonomous driving is another scenario where machines replace humans.

The Beijing Municipal Commission of Economy and Information Technology is seeking public opinions on the "Beijing Autonomous Driving Vehicle Regulations (Draft for Comments)", intending to provide clear, transparent, and predictable institutional norms for market entities of L3 and above autonomous driving vehicles while adhering to the safety bottom line. It explicitly supports the use of autonomous driving vehicles in urban public bus transportation, taxi transportation (ride-hailing), car rental, and other urban travel services.

While looking forward to the arrival of autonomous driving, there are also discussions about whether the popularization of autonomous driving will affect the employment of ride-hailing drivers. However, everything has two sides. L3 refers to conditional autonomous driving, which mainly relies on the vehicle's autonomous driving, but the main driver still has a safety officer to assist in operation. The key application scenarios for L3 are public transportation, ride-hailing, car rental, and other scenarios. But new demands are always born in new scenarios. In May 2024, the Ministry of Human Resources and Social Security announced 19 new professions. Among them, the intelligent connected vehicle tester refers to personnel who use tools, measuring instruments, testing instruments, and equipment to verify and test the functions of intelligent connected vehicles and their related components.

In the recently released "Basic Information of the Joint Body for Access and Road Traffic Pilot of Intelligent Connected Vehicles", while paying attention to which companies are selected (BYD, NIO, Changan Automobile, GAC Passenger Cars, SAIC Group, BAIC BluePark, FAW Group, SAIC Hongyan, and Yutong Bus), another thing worth noting is that the pilot products cover three major categories: passenger cars, buses, and trucks. That is to say, the L3 level is not only changing the passenger car market, but also can be imagined based on public transportation, logistics, and more possibilities.

The arrival of L3 is the starting point of autonomous driving, but it is not the end of intelligent driving.

Leave a Reply