At present, China's automotive intelligent driving system does not have autonomous driving function. What impact will it have on automotive electronic

At the press conference of the State Council Information Office on July 23, 2025, the Ministry of Public Security clearly emphasized that the "intelligent driving systems" of currently available cars in China do not have autonomous driving functions and only belong to the category of assisted driving (L0L2 level). Drivers are still the first responsible person for driving safety; If the system is allowed to operate freely (such as driving without hands or eyes), it will face three legal risks: civil, administrative, and criminal. The impact of this statement on the automotive electronics industry can be systematically summarized as follows:

1、 Technical roadmap: shifting from hardware stacking to safety redundancy and fusion perception
Rebalance between Lidar and Pure Vision Solutions
The Ministry of Public Security's clarification of the "driver's responsibility subject" has forced car companies to strengthen the safety redundancy design of their assisted driving systems. Lidar may become a key choice for car companies to improve system reliability, especially in mid to high end models, due to its centimeter level accuracy (better than pure vision) in extreme weather conditions. However, limited by cost (over 10000 yuan per unit) and power consumption, pure visual solutions still have competitiveness in the mid to low end market. The future trend is more inclined towards the fusion solution of "LiDAR+vision", which reduces the missed detection rate through multi-sensor complementarity and meets the dual needs of safety and cost.

2. Chip and algorithm optimization focus on the "human-machine co driving" scenario
The decision chip needs to optimize the real-time performance of human-computer interaction, for example:
Enhance takeover response capability (such as Horizon Journey 6 chip compressing end-to-end latency to within 120ms through BPU architecture);
Strengthen the priority of DMS driver monitoring system to ensure traceability of driver status.
At the algorithmic level, it is necessary to reduce reliance on long tail scenarios, such as through reinforcement learning and world model simulation (such as Horizon Senna architecture), to replace some real vehicle testing and accelerate iteration in response to extreme working conditions.

2、 Legal support: Responsibility definition drives standardization of electronic systems
1. Strong supervision of car companies' promotion and user training
The Ministry of Public Security requires car companies to eliminate exaggerated propaganda (such as "off hand driving" and "AI assisted driving" terms), and mandates new users to train on the boundaries of assisted driving functions. This leads to:
Real time functional status prompts (such as system activation/exit, responsibility switching reminders) need to be added to the car machine interaction interface;
Electronic manuals and training videos need to pass national certification standards, and some car companies have urgently modified their promotional materials (such as Xiaomi and Xiaopeng deleting the description of "advanced intelligent driving").

2. Electronic system for driving test and regulatory adaptation
The Ministry of Public Security plans to include the operation standards of assisted driving in the scope of driving tests, which means:
The simulated driver needs to integrate an L2 level system operation module;
The electronic architecture of vehicles needs to open standardized data interfaces for traffic management departments to supervise functional compliance (such as whether it is mandatory to hold the steering wheel for detection).

3、 Industrial pattern: supply chain restructuring and accelerated domestic substitution
1. Upstream electronic components: Safety certification becomes the core threshold
Perception layer: Lidar manufacturers (such as Sagitar Juchuang and Hesai Technology) need to obtain automotive grade functional safety certification (ISO 8800) and optimize costs (target $500/piece) to match scale requirements;
Decision makers: Chip companies need to break through the "computational efficiency cost" triangle contradiction (such as Horizon BPU architecture improving energy efficiency by 40%), while meeting the expected functional safety (SOTIF) standards;
Execution layer: The wire controlled chassis supplier needs to implement dual redundancy design (response<120ms) to ensure backup reliability of braking/steering.

2. Midstream car companies: shifting from functional competition to systematic capability
The core of industry competition has shifted from hardware parameters (such as TOPS computing power) to:
Data closed-loop capability: Building a digital twin testing system with over 100000 scene libraries;
Accumulation of vehicle safety regulations: Horizon Robotics has verified through mass production of 8 million chips, covering 90% of China's road scenarios.
By 2025, the L2+penetration rate will reach 30%, but the willingness of users to pay will decrease to 28%, making cost control key (such as the NOA in urban areas dropping to 150000 yuan for car models).

4、 Consumers and Markets: Cognitive Correction and Demand Differentiation
1. Responsibility risk education to reduce abuse rate
The Ministry of Public Security has issued a triple legal liability warning (civil compensation, revocation of driver's license, criminal liability), combined with propaganda emphasizing that "taking off hands and eyes=high-risk behavior", which will significantly reduce users' excessive dependence on assisted driving. Car companies need to enhance their real-time warning capabilities through electronic systems such as OTA upgrades.

2. Functional requirements are centralized towards basic ADAS
Market research shows that consumer demand for basic ADAS features such as AEB automatic emergency braking remains stable at over 75%, far higher than L2+payment willingness (28%). This promotes the sinking of electronic solutions, such as the simplified configuration of 11 cameras and 1 millimeter wave radar, replacing high cost laser radar solutions.

5、 Future trend: Regulations and technology collaborate to promote the implementation of L3+
Clarifying legal attributes: The Ministry of Public Security is promoting the revision of the Road Traffic Safety Law, defining the responsibility for L3 level "human-machine co driving" accidents (with Beijing pilot car companies taking the main responsibility), paving the way for advanced autonomous driving;
Vehicle road cloud collaboration has become a breakthrough point: the current V2X coverage rate is less than 30%, and it is necessary to support the L4 level operation target in 50 cities by 2025 through 5G latency compression (<1ms) and Beidou high-precision positioning (± 5mm).

Conclusion: The automotive electronics industry is facing a "redefinition of safety"
The statement from the Ministry of Public Security marks the official entry of the industry into the "compliant competition" stage: in the short term, electronic systems need to strengthen human-computer interaction and responsibility tracing functions, and bear the cost of promoting compliance; In the long run, technology will develop towards a dual track of "safety redundancy+experience penetration", promoting domestic chip/sensor manufacturers to seize the incremental market through systematic capabilities (vehicle regulation certification, data closed-loop), and ultimately supporting the landing of L4 level autonomous driving within the responsibility framework.

The article material is sourced from the internet. If there is any infringement, please contact us for deletion: marktian@uxigroup.com