国内新闻 我国科学家取得锂电池核心技术首创性突破 2月26日，据央视新闻报道，近日，由南开大学和上海空间电源研究所等单位科研人员组成的团队，取得了一项首创性的突 破。科研团队突破了氟难以溶解锂盐等关键难题，合成出系列新型氟代烃溶剂分子，通过调控氟原子的电子密度和溶剂分子的 空间位阻，既显著降低电解液用量，又具有快速电荷转移的动力学特性，从而同时提升了电池能量密度和低温适应能力。这一 成果在国际学术期刊《自然》上发表。 商务部：中德双方企业达成涉及汽车等行业十余项商业协议 2月26日，商务部举行例行新闻发布会。商务部新闻发言人何咏前在回答有关德国总理默茨在中国访问的问题时表示，中德 双方在经贸领域开展了深入交流，取得了积极务实成果。访问期间，双方企业达成了十余项商业协议，涉及汽车、机械、能 源、物流、金融等行业。 默茨在奔驰中国总部体验中德合作的辅助驾驶系统 2月26日，正在中国进行正式访问的德国总理默茨在位于北京的奔驰中国总部试乘了新一代S级轿车，体验由梅赛德斯-奔驰 与中国科技公司Momenta合作开发的城区及高速领航辅助驾驶系统。他对此评价道："令人惊叹。"在参访过程中，默茨总理还了 解了梅赛德斯-奔驰在 ...

这套系统能根据线路坡道和信号条件自动计算并调整车速，就像汽车的辅助驾驶系统一样，在平直道上 自动维持速度，遇到上坡会自动加大电流，下坡则适时制动。"以前开车，手柄得一直攥着，精神高度 紧张，一趟跑下来肩膀都是僵的。"王祖贵说，"现在辅助驾驶系统能完成大部分重复操作，我们能更专 心地看线路、做联控，轻松不少。" "列尾风压正常，机班人员到位，具备发车条件，动车。"2月23日20时05分，随着一声长鸣，王祖贵推 动手柄，驾驶一列满载万吨煤炭的重载列车，缓缓驶离国家能源集团新朔铁路点岱沟车站，向着神池南 站进发。 记者此时就站在王祖贵身后，目光落在他面前的仪表盘上，那些数字正不停闪烁跳动。"咱们这列车一 共有108节车厢，全长1.35公里。"王祖贵一边说着，一边习惯性地扫视着前方，"从车头走到车尾，最 快也得10多分钟。" 41岁的王祖贵是新朔铁路机务公司的一名司机，也是这条煤炭运输大动脉上的"老熟人"，今年是他在重 载铁道线上奔走的第17个年头。 王祖贵的眼睛盯着前方的同时，也会不时扫一眼面前的仪表盘。上面跳动的，除了车速、管压这些传统 数据，还有一些他以前想都不敢想的东西——辅助驾驶系统状态指示灯。 在王祖贵头 ...

经查，王某群所驾汽车安装有2级驾驶自动化系统，亦即辅助驾驶系统，具有辅助驾驶功能。该系统设 定，若驾驶人双手脱离方向盘超过2分钟，系统会提示驾驶人手握方向盘、接管车辆，若未及时接管， 车辆会主动减速并退出系统。 2026年2月13日，最高人民法院首次发布道路交通安全刑事专题指导性案例，《中国经营报》记者注意 到，该批案例涵盖以危险方法危害公共安全、交通肇事、危险驾驶三类案件。 随着辅助驾驶技术的普及应用，也出现了诸多违规乱象：有的驾驶人激活辅助驾驶系统后不再专注驾 驶，转而玩手机、睡觉；更有甚者购买、使用"智驾神器"等非法配件，逃避系统安全监测，长时间"脱 手"驾驶，严重威胁道路交通安全，也给案件的刑事责任认定规则带来争议。 指导性案例271号《王某群危险驾驶案》，专门明确激活辅助驾驶功能情形下驾驶人的刑事责任认定规 则：车载辅助驾驶系统不能代替驾驶人成为驾驶主体，驾驶人激活车载辅助驾驶功能后，仍是实际执行 驾驶任务的人，负有确保行车安全的责任。 该案例进一步明确：行为人激活辅助驾驶功能，并利用私自安装的配件逃避辅助驾驶系统监测的，即使 其不在主驾驶位实际操控机动车，仍应作为驾驶主体承担相应法律责任。 案情显 ...

Core Insights - Mercedes-Benz reported a significant decline in financial performance for 2025, with revenue dropping to €132.21 billion, a 9.2% decrease from €145.6 billion in 2024 [1][3] - The adjusted EBIT fell to €8.2 billion, down from €13.7 billion in 2024, while net profit plummeted by 48.8% to €5.33 billion [1][3] - The company anticipates that revenue for 2026 will remain at the same level as in 2025 [3] Financial Performance - Revenue for 2025: €132.21 billion, down 9.2% from €145.6 billion in 2024 [1][3] - Adjusted EBIT for 2025: €8.2 billion, compared to €13.7 billion in 2024 [1][3] - Net profit for 2025: €5.33 billion, a decrease of 48.8% from the previous year [1][3] - Free cash flow from industrial operations: €5.4 billion, down from €9.2 billion in 2024 [1][3] - Net working capital: €32.2 billion, slightly up from €31.4 billion at the end of 2024 [1] Sales Performance - Total sales for 2025: 2.16 million units, a 10% decline year-on-year [3] - Sales of pure electric vehicles: 197,300 units, down 4% [3] - Sales of Mercedes-Benz cars: 1.8 million units, down 9% [3] - Sales of Mercedes-Benz vans: 359,100 units, down 11%, with electric van sales increasing by 46% to 28,500 units [3] - Regional sales: Europe at 634,600 units (down 1%), North America at 320,600 units (down 12%), and China at 551,900 units (down 19%) [3] Strategic Initiatives - The company is focusing on a significant technological and product push, planning to launch over 40 new models within three years [3] - In China, Mercedes-Benz achieved over 2.4 million OTA upgrades in 2025 and has implemented advanced driver assistance systems across 17 core models [4] - Collaboration with ByteDance has led to a 97% active usage rate of the AI-powered MBUX virtual assistant [4] - A partnership with Momenta has resulted in the rollout of advanced driver assistance systems, with 9 new models set to feature these systems in 2026 [4]

Core Insights - The autonomous driving technology company, Qingtian Zhihang, announced that its assisted driving systems have surpassed 1 million units in deployment, marking a significant milestone for the company and the industry [1][2] - The CEO, Yu Qian, stated that 2026 will be a pivotal year, marking the beginning of a decade of advancements in autonomous driving, transitioning from "human-like intelligence" to "superhuman intelligence" [1] Industry Developments - The development of autonomous driving technology is categorized into three phases: 1. "Machine Intelligence" from 2004 to 2020, characterized by weak generalization capabilities 2. "Human-like Intelligence" from 2021 to 2025, where systems approach human driving levels 3. The onset of "Superhuman Intelligence" in 2026, where systems are expected to exceed human driving capabilities [1] - The introduction of end-to-end architectures and models like VLA and world models will enhance the understanding of physical laws, human intentions, and experiential knowledge, leading to a safety level ten times that of human drivers [1] Future Projections - Qingtian Zhihang outlined eight industry predictions for the next five years: 1. Equal access to intelligent driving with urban navigation assistance becoming standard in domestic vehicles 2. Significant improvements in technology reliability, with urban NOA takeover frequency dropping to a monthly basis 3. A restructured insurance system with autonomous driving premiums being over 50% lower than human driving premiums 4. Robotaxi services operating 24/7 in 50 global cities, with daily orders surpassing human ride-hailing services 5. The emergence of true L4 level vehicles without a driver's cabin entering the personal consumer market 6. Global autonomous logistics vehicles (Robovans) exceeding 2 million units 7. Enhanced travel safety experiences for children, the elderly, and disabled individuals 8. The arrival of a general physical AI era, with intelligent robots becoming part of daily life and work [1] Company Strategy - Qingtian Zhihang has adopted a dual strategy of mass-producing L2++ systems while concurrently developing L4 technologies, supported by a data closed-loop from an "autonomous driving super factory" [2] - The company has established deep collaborations with nearly 10 major domestic automakers, aiding in the mass production of 23 vehicle models, with over 50 new models expected to launch in 2026, all equipped with urban NOA [2] - The recent launch of the first urban NOA solution based on Horizon Journey 6M has been integrated into the Ideal L series smart vehicle [2]

Investment Rating - The report maintains a "Buy" investment rating for the company [9] Core Views - The company is experiencing high growth, with a projected net profit of approximately 3.85 billion yuan for 2025, reflecting a year-on-year increase of 30% [5] - The penetration of advanced driver-assistance systems is accelerating, with the company leveraging high-value products to expand its customer base and increase revenue and profit [6] - The company is focusing on the development of autonomous buses, with significant progress in the Robobus project and a dedicated testing facility for L4 autonomous driving capabilities [7] - Revenue forecasts for 2025-2027 are 26.7 billion, 33.5 billion, and 41.5 billion yuan respectively, with corresponding EPS of 2.19, 2.95, and 3.67 yuan, indicating a strong growth trajectory [8] Summary by Sections Market Performance - The company has shown significant improvement in Q4, with a projected net profit of about 1.11 billion yuan, a substantial increase from 0.71 billion yuan in Q3 [5] Financial Forecasts - The company is expected to achieve a net profit growth rate of 36.3% in 2025, 34.6% in 2026, and 24.6% in 2027, with a projected ROE of 20.5% in 2025 and increasing to 26.1% by 2027 [11][14]

Investment Rating - The report maintains a "Buy" investment rating for the company [9] Core Insights - The company is experiencing high growth, with a projected net profit of approximately 3.85 billion yuan for 2025, reflecting a year-on-year increase of 30% [5] - The penetration of advanced driver-assistance systems is accelerating, with the company leveraging high-value products to expand its customer base and increase revenue and profit [6] - The company is focusing on the development of autonomous buses, particularly the Robobus project, which aims to enhance its automatic driving capabilities and reduce the time from R&D to application [7] Financial Performance - For the first three quarters of 2025, the company achieved a net profit of 274 million yuan, with an expected significant increase in Q4 to approximately 1.11 billion yuan, marking a substantial quarter-on-quarter improvement [5] - Revenue forecasts for 2025-2027 are projected at 26.7 billion, 33.5 billion, and 41.5 billion yuan respectively, with earnings per share (EPS) expected to be 2.19, 2.95, and 3.67 yuan [8] - The company is expected to maintain a gross margin of over 60% from its new AI model, contributing to profit growth in 2026 [6] Market Position - The company is positioned as a leader in commercial vehicle intelligent solutions, with significant growth potential in the autonomous driving sector [4] - The report highlights the company's strategic initiatives in expanding its market opportunities in various scenarios, including public transport and logistics [7]

Group 1 - By the end of 2025, China's first batch of L3 autonomous vehicles will receive exclusive licenses, marking the transition from technical testing to commercial deployment [2] - In early January 2026, Yuanrong Qixing announced a partnership with a leading international OEM for L3 autonomous driving, representing the first international collaboration of its kind for domestic smart driving companies in 2026 [2] - The introduction of exclusive licenses and international collaboration is seen as a significant milestone, indicating that 2026 will be the year when L3 autonomous driving commercialization truly begins [2] Group 2 - L3 autonomous driving represents a turning point, shifting from driver-assisted systems (L2 and L2+) to highly automated systems where the vehicle takes over driving tasks under specific conditions [3] - The transition to L3 requires higher safety and reliability standards, focusing on environmental perception stability and decision-making predictability [3] - The challenge in moving from L2+ to L3 lies in redefining responsibility; once a system is classified as L3, the responsibility shifts from the driver to the system during its operation [4] Group 3 - Yuanrong Qixing's collaboration with an international OEM is a milestone in China's autonomous driving sector, moving beyond previous partnerships that focused on Robotaxi demonstrations and hardware collaborations [4] - Yuanrong Qixing aims for L4 fully autonomous driving from the outset, emphasizing the system's understanding and judgment capabilities without relying on high-precision maps [5] - The company's city NOA system is designed for nationwide deployment without the need for city-specific adaptations, showcasing its technological advancements [5] Group 4 - In 2025, over 200,000 vehicles equipped with Yuanrong Qixing's assisted driving system are expected to be delivered, involving more than 15 mass-produced models [6] - The collaboration with domestic OEMs allows for faster project rhythms and shorter feature rollout cycles, necessitating efficient communication and delivery mechanisms [6] - The differences in focus between domestic and international OEMs highlight varying approaches to L3 product planning, with domestic firms prioritizing efficiency and rapid iteration, while international firms emphasize safety and regulatory compliance [6]

Core Viewpoint - The decline of independent intelligent driving companies incubated by traditional automakers, such as Haomo Zhixing, highlights the challenges and inefficiencies in the current market, leading to a shift towards more collaborative and cost-effective solutions [1][10][13]. Group 1: Haomo Zhixing's Decline - Haomo Zhixing announced a complete work stoppage starting November 24, 2024, due to its operational status, with indications of significant layoffs and management departures since 2024 [3][4]. - The company, which originated from Great Wall Motors in 2019, had once been a leading player in the intelligent driving sector, achieving a valuation exceeding $1 billion and planning an IPO in 2025 [4][5]. - Despite initial successes, Haomo Zhixing's actual deployment of its NOH system was limited to only three cities by the end of 2023, falling behind competitors [5][6]. Group 2: Industry Trends and Challenges - The trend of traditional automakers dissolving or integrating their independent intelligent driving subsidiaries, such as the dissolution of Dazhuo Intelligent and the integration of Zero束 Technology into SAIC, reflects a broader industry shift towards consolidation and collaboration [7][8]. - The low return on investment for these independent companies has led automakers to reconsider their support, especially in light of more cost-effective third-party solutions [10][11]. - The competitive landscape has evolved, with over 80% of domestic automakers opting for partnerships with established tech firms like Huawei, indicating a preference for collaborative models over independent development [14][15]. Group 3: Future Directions - The industry is moving towards a hybrid model of "self-research + cooperation," allowing traditional automakers to leverage their manufacturing strengths while addressing technological gaps through partnerships [15]. - The market is increasingly favoring third-party intelligent driving solutions, as evidenced by the dominance of companies like Momenta and Huawei in the NOA market, which has shifted the focus back to core automotive manufacturing [14][15].

Core Insights - The article discusses the transition of AI from model capability competition to execution capability in the physical world, highlighting the challenges and opportunities in this domain [2][3]. Company Summaries - Zhi Bian Liang is focused on developing embodied intelligence foundational models and general-purpose robots, emphasizing the need for a physical model that operates in the real world, distinct from virtual models [4]. - Yuan Rong Qi Xing has been involved in autonomous driving, witnessing the industry's evolution from high-precision mapping to end-to-end models, and has successfully deployed 200,000 vehicles with their driving assistance systems, with a projection of reaching one million vehicles next year [5]. Challenges in AI Implementation - The transition from simulation to real-world application presents significant challenges, including the need for extensive pre-training based on real-world data, which is not easily replicated in simulated environments [6][7]. - The physical world introduces complexities that are not present in simulations, such as the need for precise manipulation and the impact of minor errors on outcomes [9][10]. Importance of Data and Training - The collection of vast amounts of real-world data is crucial for effective pre-training, and the integration of language models can enhance learning efficiency [7][18]. - The current data generation from 200,000 vehicles is substantial, necessitating careful selection and quality control to optimize model performance [18]. Future of Commercialization - The commercialization of embodied intelligence is expected to gain momentum by 2026, with predictions of significant advancements in practical applications and return on investment [21][22]. - The industry is currently in a phase similar to early autonomous driving, with many companies still in the demo stage, but there is optimism about achieving scalable commercial applications soon [19][20]. Role of Language Models - Language models are seen as essential for providing supervisory information during training, aiding in the rapid learning of complex tasks [12][13]. - However, there is debate about the necessity of language in physical AI, with some arguing that while it enhances understanding, it may not be critical for all applications [15][26]. Technical Considerations - The development of physical AI models requires overcoming significant engineering challenges, including the need for real-time feedback and the limitations of current computational resources [25][26]. - The scaling laws in AI suggest that with sufficient data and resources, it is feasible to train models that can operate effectively in the physical world within a reasonable timeframe [24][26].