L3车型安全兜底方不一定是单一主体 2025年12月，工信部宣布长安汽车（000625）旗下深蓝汽车品牌、北汽极狐获得我国首批L3级有条件 自动驾驶车型准入许可，我国自动驾驶汽车产业正式从"技术验证"迈向"量产应用"新阶段。 "智驱新程·芯动未来"新京报第二十届超级汽车论坛1月30日举行。深蓝汽车软件开发总监余龙在在"自 动驾驶挺进L3，带来怎样的改变？"圆桌对话环节中表示，L2级跨越到L3级，本质是辅助驾驶到有条件 自动驾驶的突破，是责任主体的转移。他预计，2026年，L3级将在To B运营端快速推进；L2级城区领 航辅助系统将快速渗透，有望渗透到15万级别车型。 余龙表示，从L2级跨越到L3级有三个关键点，一是系统可靠性，二是功能安全和预期功能安全的提 升，三是人机接管的边界，四是运行设计域精细化管理。 "L2跨越到L3的本质是辅助驾驶到有条件自动驾驶的突破，意味着责任主体的转移。"余龙解释称，L2 阶段驾驶员是责任主体，L3阶段责任主体可能是主机厂、系统供应商等。技术难点在于责任主体转移 过后，如何把风险控制到最低。此外，AI（人工智能）决策模型的确立性和可解释性需进一步突破。 在他看来，L3车型不一定 ...

Core Viewpoint - Safety remains the most important topic in the automotive industry and is the starting point for smart vehicles and the ultimate test for autonomous driving [1] Group 1: Autonomous Driving Levels and Regulations - The global automotive automation level classification is consistent, ranging from L0 to L5, with L2 being entry-level assisted driving and L3, L4, and L5 classified as autonomous driving [3] - In December 2025, the Ministry of Industry and Information Technology announced that Changan Automobile and BAIC Jihe received the first batch of L3 conditional autonomous driving vehicle permits in China [3] - The penetration rate of vehicles with L2 combined driving assistance functions is expected to reach 66.1% by the end of 2025 [4] Group 2: Limitations of OTA Upgrades - It is not feasible to upgrade L2 vehicles to L3 through OTA (Over-The-Air) updates due to hardware limitations [4] - The electronic architecture of L2 vehicles does not meet the safety and computational requirements for L3 autonomous driving [4] Group 3: Role of AI in Smart Vehicles - AI is driving the revolution in automotive intelligence and is considered the "soul" of smart electric vehicles, with electric vehicles serving as the foundational platform [5] - The use of language models in smart cockpits and the integration of deep learning models in intelligent driving are becoming common [6] - By 2026, visual language models and world models are expected to be significant trends in intelligent driving [6] Group 4: Safety Challenges in Autonomous Driving - Key safety issues remain unresolved in autonomous driving, which is a major reason for its lack of realization [8] - Unknown risks during the operation of combined assistance driving systems include dangerous takeovers due to misidentification of targets and emergency system misactivations [9] - The requirements for safety in autonomous driving have expanded to include cybersecurity, functional safety, and expected functional safety [9] Group 5: Future of Smart Vehicles - The development of smart vehicles has reached a crossroads, with L2 intelligent assisted driving improving vehicles, L3 autonomous driving potentially revolutionizing them, and L4 driverless technology transforming mobility [10] - There is an expectation for L3 vehicles to achieve stable operation at full speed on highways, including speeds of 0-120 km/h and above [10]

Core Insights - The latest data from the China Association of Automobile Manufacturers indicates that by 2025, China's automobile production and sales are expected to reach 34.53 million and 34.40 million units, respectively, representing year-on-year growth of 10.4% and 9.4%, maintaining a scale of over 30 million units for three consecutive years and ranking first globally for 17 years [1] Industry Overview - The combination of policies such as trade-in programs for consumer goods, the promotion of new energy vehicles in rural areas, and enhanced supply-demand adaptability is driving the sustained release of demand potential in China's automobile market [1] - The automobile manufacturing industry has a long supply chain and significant employment absorption, playing a notable role in economic growth [1] New Energy Vehicles - By 2025, the production and sales of new energy vehicles in China are projected to reach 16.626 million and 16.49 million units, respectively, with year-on-year growth of 29% and 28.2%, maintaining the global lead for 11 consecutive years [1] - New energy vehicles are expected to account for 47.9% of total new car sales [1] Technological Innovations - Key core technologies such as next-generation power batteries, electronic and electrical architectures, smart cockpits, and intelligent driving assistance are experiencing continuous breakthroughs [1] - By 2025, the mass application of 800-volt high-voltage silicon carbide platforms and fast-charging technology that allows for 80% charge in 15 minutes is anticipated, along with the approval of the first batch of L3-level conditional autonomous driving vehicles [1]

2026年开年，全球自动驾驶领域在技术验证与商业化探索上继续前行。我国首批L3级有条件自动驾驶 车型的准入许可，特斯拉将在二季度启动无方向盘、无踏板车型量产的计划，以及Waymo持续扩张无 人出租车服务网络，标志着自动驾驶的竞争重点，正从攻克"能否实现"的技术挑战，转向如何构建一个 可盈利、安全且被广泛接受的生态系统。 近日，麦肯锡未来移动中心发布了第三版自动驾驶行业领导者调查报告。这份基于全球行业专家洞察的 报告揭示了一个关键趋势：在经历了早期狂热之后，自动驾驶行业正褪去光环，进入一个更加务实、理 性且充满复杂挑战的新发展阶段。 时间表普遍延后：商业化预期趋于保守 报告最直观的发现是，行业对自动驾驶普及速度的预期显著回调。与2023年预测相比，2025年度调查报 告将大多数自动驾驶应用场景的普及时间表平均推迟1-2年。尽管中美部分城市已开通L4级无人驾驶出 租车服务，但全球性大规模商业化推广的预期节点还是从2029年推迟至2030年。面向私人乘用车的L4 级城市试点预计将从2030年推迟至2032年，而完全自动驾驶卡车的商业可行性也从2031年延后至2032 年。总体而言，专家认为L4级自动驾驶将首先在出租 ...

Core Insights - The article emphasizes the importance of collaborative and integrated approaches in building new scenarios, as highlighted by the government's implementation opinions supporting the construction of "integrated scenarios in industry fields" [1] Group 1: Industry Developments - The establishment of demonstration zones for comprehensive road testing relies not only on advancements in individual technologies but also on the integration and promotion of new technologies, enhancement of the entire industry chain's innovation capabilities, precise policy support, and a favorable collaborative innovation ecosystem [1] - A comprehensive effect is formed through the path of "technological breakthroughs—scenario validation—industry application—system upgrades," which aids in the realization of new scenarios and maximizes their industrial driving effects [1] Group 2: Regulatory Approvals - The Ministry of Industry and Information Technology recently announced the first batch of L3-level conditional autonomous driving vehicle approvals, allowing related models to conduct pilot tests in designated areas of Beijing and Chongqing [1] - This development indicates that applications of intelligent connected vehicles and "vehicle-road-cloud integration" are gradually maturing, suggesting a broader future outlook for these technologies [1]

Core Insights - The commercialization of autonomous driving is advancing with the recent approval of two L3 conditional autonomous driving models in China, marking a significant transition from technology validation to mass production application [1][2] - L3 level corresponds to conditional automation where the driving task can be delegated to the automated system under specific conditions, representing a critical threshold from assisted driving to autonomous driving [1][2] - The approval of L3 models indicates that their safety operation capabilities have undergone preliminary verification, necessitating updates in traffic accident liability, insurance product design, and human-machine co-driving ethical standards [1][2] Group 1: Industry Development - The development of autonomous driving technology is not a solitary technical advancement but involves integrating safety as a fundamental aspect of industry evolution, adhering to the principle of "safety first, gradual progress" [2][3] - The path of development includes the initial deployment of function-specific unmanned vehicles in logistics and delivery, gradually moving towards cautious testing of passenger-carrying autonomous vehicles in more complex environments [2][3] - As autonomous driving technology matures, the transition from closed to open roads indicates an increasing proximity to large-scale application [2][3] Group 2: Safety and Responsibility - The autonomous driving sector has a very low tolerance for error; without a solid safety foundation, any efficiency gains or innovations could be illusory [3] - Public trust in autonomous driving is fragile and can be easily undermined by incidents such as system failures or misrepresentations of technology capabilities [3] - Establishing a robust safety framework and clarifying responsibility boundaries among manufacturers, software providers, and drivers are essential for the stable and orderly development of the industry [2][3] Group 3: Future Outlook - The current advancements in autonomous driving are opening up significant commercial opportunities and will greatly impact people's travel and daily lives [3] - Companies that have received approval emphasize that technical approval is just the starting point, with the ultimate goal being the safe return of passengers [3] - The industry has accumulated experience and advantages in technology innovation and safety regulation, suggesting a patient and steady approach is necessary for building a reliable safety infrastructure [3]

Core Insights - The commercialization of autonomous driving has taken a significant step forward with the recent approval of two L3 conditional autonomous driving models in China, marking a transition from technology validation to mass production application [1][2] - L3 level autonomous driving allows the automated system to take over driving tasks under specific conditions, representing a critical milestone in the evolution from assisted driving to fully autonomous driving [1][2] - The approval of L3 models indicates that their safety operation capabilities have been preliminarily validated, necessitating updates in traffic accident liability, insurance product design, and human-machine interaction ethics [1][2] Industry Development - The development of autonomous driving technology is guided by the principle of "safety first, gradual progress," with initial applications in logistics and delivery before moving to more complex environments [2][3] - As autonomous driving technology matures, the industry is approaching large-scale application, highlighting the need to validate safety redundancy mechanisms in extreme scenarios and clarify responsibility boundaries between humans and machines [2][3] - The industry faces challenges such as cost control and technology upgrades, but the focus remains on ensuring safety as the core value of autonomous driving [2][3] Public Trust and Safety - The tolerance for errors in the autonomous driving sector is extremely low, making public trust crucial yet fragile; any incidents can significantly undermine confidence and halt commercialization efforts [3] - Companies emphasize that obtaining technical approval is just the beginning, with the ultimate goal being the safety of passengers, which is essential for gaining public trust [3] - The industry is encouraged to maintain patience and focus on building a solid safety foundation to ensure that every journey is trustworthy, thereby expanding the potential for autonomous driving to enhance societal welfare [3]

Group 1 - The recent approval of the first batch of L3-level conditional autonomous driving vehicles marks a significant step towards commercialization in China, transitioning from technical validation to conditional road use [1] - The core breakthrough of "conditional road use" lies in balancing institutional innovation with risk control, clearly defining responsibility during system takeover between drivers and car manufacturers [1] - The launch of the "conditional road use" era is a natural outcome of industry development, with the penetration rate of L2-level assisted driving in new vehicles reaching 64%, indicating increasing consumer acceptance of intelligent driving [1] Group 2 - The entry into the "conditional road use" phase will unleash significant economic value and reshape the automotive industry ecosystem, benefiting upstream core component manufacturers and driving down costs through mass production [2] - Automotive companies will shift from hardware sales to a "hardware + software + service" profit model, leveraging services like OTA upgrades to extract long-term value [2] - The maturity of autonomous driving technology supports the transformation of automotive business models from one-time vehicle sales to full lifecycle service operations, positioning autonomous vehicles as "mobile intelligent terminals" [2]

Core Viewpoint - The issuance of the first L3-level highway autonomous vehicle special license plates in Beijing marks a significant milestone in China's autonomous driving development, transitioning from technical validation to mass production application [1][4]. Group 1: L3-Level Autonomous Driving Overview - L3-level autonomous driving represents a critical transition from "driving assistance" to "autonomous driving," allowing the system to perform driving tasks under certain conditions without continuous driver monitoring [2]. - The first three quarters of this year saw a 21.2% year-on-year increase in new car sales with L2-level driving assistance features, achieving a penetration rate of 64% [2]. Group 2: Regulatory Framework and Testing - The Ministry of Industry and Information Technology (MIIT) has issued strict limitations on the operational scenarios, road types, urban areas, and speed limits for the two approved models, emphasizing a cautious regulatory approach [3]. - There is a distinction between "road test licenses" issued by local authorities and "product access licenses" from national departments, with the latter being significantly more challenging to obtain [3]. Group 3: Future Implications and Challenges - The successful implementation of L3-level autonomous driving requires not only technological advancements but also supportive policies, industry ecosystems, and infrastructure [4]. - The current pilot approach is characterized by "small-scale initiation and conditional implementation," indicating that mass production is not imminent despite the recent approvals [4].

Core Insights - The Ministry of Industry and Information Technology has officially granted the first batch of L3 conditional autonomous driving vehicle permits in China, consisting of two models, indicating a significant step towards advanced autonomous driving technology [1] Group 1: L3 Level Autonomous Driving - The transition from L2 to L3 level represents a revolutionary change, as L3 is classified as autonomous driving rather than merely assisted driving, with the system taking primary responsibility during operation [2] - The first approved models are purely electric vehicles, as they are easier to integrate with intelligent driving systems and have fewer control links [2] Group 2: Safety and Regulations - The permitted usage of L3 vehicles is subject to strict limitations on specific road segments, with speed limits set at 50 km/h in Chongqing and 80 km/h in Beijing, reflecting a cautious approach to safety [3] - Clear responsibility boundaries have been established, where the system assumes liability in the event of accidents during autonomous operation, addressing previous consumer misunderstandings regarding assisted driving [3] - Supporting regulations, such as the "Automated Driving Vehicle Management Regulations" set to be implemented by the end of 2024 in Beijing, and a dedicated insurance system with a maximum payout of 5 million yuan per incident, have been put in place [3] Group 3: Industry Transformation - The introduction of L3 technology is expected to shift the automotive industry from a focus on selling vehicles to providing safe mobility services, with competition evolving from price wars to safety and responsibility capabilities [6] - A complete ecosystem involving automakers, technology companies, insurance firms, and regulatory bodies is necessary for the successful implementation of L3 autonomous driving [6] - Future advancements in autonomous driving are anticipated to drive transformations in road collaboration and smart city development, with initial trials of smart roads already underway [6] Group 4: Future Outlook - Despite the approval of L3 vehicles, experts believe that higher levels of autonomous driving will require a gradual approach, with full realization of L4 potentially taking five to ten years, necessitating synchronized advancements in regulations, infrastructure, and user education [7] - The approval of the first L3 conditional autonomous driving models not only marks a technological breakthrough but also signifies a profound restructuring of responsibility, ecosystem, and urban forms [7]