1.How Will Autonomous Driving Evolve?

Technological progress drives the continuous upgrading of autonomous driving functions from low to high levels, bringing an increase in per-vehicle value. 2015 Before, driver assistance functions were mainly at L1/L0 level, with representative functions being LKA, AEB , etc. The automotive E/E architecture was distributed, with per-vehicle value of approximately 7 thousand yuan. 2016 entered the L2 level era, with representative functions being ACC with LKA, APA, some ECU began integrated development, but there was still no domain division, with per-vehicle value within 1.5 thousand yuan. By 2020 it will formally enter the L3 level era, with representative functions being TJP, RPK , etc. The distributed architecture gradually evolved into a domain-centralized architecture, with per-vehicle value of approximately 2.5 thousand yuan. By 2023 it will gradually enter the L4 level era, with representative functions being City Pilot , etc., with per-vehicle value exceeding 4 thousand yuan.

2.How Large is the Autonomous Driving Market?

Based on the demand for improved efficiency +and safety, consumers are willing to pay, driving downstream OEMs to actively increase ADAS penetration rates. Based on collecting 118 star sample vehicle models' major ADAS functions' 2017-2020 penetration rates, based on market capacity/ADAS function penetration rate/ per-vehicle value three-step forecast, we estimate that 2020 the domestic autonomous driving market size was 844 billion yuan, up 19.3% year-on-year, and in 2025 it will reach 2250 billion yuan, with a compound growth rate of 21.3%. The future main incremental markets will be decision-layer computing platforms, new sensors, and software algorithms. Without considering L3/L4 level autonomous driving new functions, from 2019 to 2025 the top three ADAS functions contributing to market growth (2025 market size/CAGR) are: automatic parking (384 billion yuan/+28.7%),

adaptive cruise control (312 billion yuan/+20.9%), 360-degree panoramic imaging (157 billion yuan/+10.8%).

3.Where Exactly is the Core Position of Autonomous Driving?

The automotive E/E architecture upgrade is the core of autonomous driving development , mainly reflected in:1) Hardware architecture upgrade. Development from distributed to domain-centralized controllers/central computing platform development, with the benefits of: higher computing power utilization, reducing total computing power design demand;

Unified interaction, achieving vehicle function coordination; shortening wiring harnesses, reducing weight. 2) Software architecture upgrade. Software-hardware decoupling, achieving software-defined vehicles, with the benefits of: enabling software/firmware online upgrades, soft real-time software architecture, operating system portability; collected data information can be reused,effectively reducing hardware demand.3) Communication architecture upgrade. LIN/CAN bus developing towards Ethernet, with the benefit of: meeting high-speed transmission requirements. In addition, E/E The new development of the architecture can achieve

ECU underlying standards/protocol definitions, enabling OEMs to control core technologies, and on the other hand achieving reduced development, hardware, and installation costs. Tesla has taken the lead in this E/E architecture upgrade.

4.How Will the Autonomous Driving Competitive Landscape Be Reshaped?

The existing autonomous driving industry chain is divided into three major segments: upstream (Tier2 sensors/chips/software, etc.) - midstream (Tier1 individual ADAS functions) - downstream (Tier0.5 solution integrators). Compared with traditional driving, autonomous driving adds the Tier0.5 role

, with software replacing hardware as the core barrier, leading to rapid increases in per-vehicle value and reshaping of the competitive landscape. The marginal change ranking of value: Tier0.5> Tier2> Tier1. OEMs/traditional strong Tier 1/hard tech/internet and many other core participants are increasing R&D investment, entering the autonomous driving Tier 0.5 battle. Tier 0.5 integrator competition is actually a E/E architecture definition competition. In the short term, the E/E architecture is not yet mature, with OEMs and strong Tier 1 suppliers having more obvious advantages, while in the long term

hard tech and OEMs may be the ultimate competitors.

5. Who Can Benefit from the Autonomous Driving Industry Dividend?

2025 will see an autonomous driving industry dividend exceeding 2 thousand billion yuan scale, with beneficiaries: 1) In the traditional industry chain, those relying on product upgrades or new business expansion to increase per-vehicle value, benefiting stocks [Desay SV + Xingyu Shares + NavInfo + Bethel + Joyson Electronics +

Keboda]. 2) Hard tech - internet giants leveraging advantages in chips/software algorithms/communications to enter the autonomous driving field, benefiting stocks [Huawei + Baidu + Alibaba + Tencent]. 3) OEMs improving product customer experience through autonomous driving upgrades, thereby gaining higher market share.