报告分析认为，中美引领自动驾驶产业发展的格局已形成。两国在自动驾驶领域已形成了"技术-产 业-场景-监管"四位一体的领先发展模式，为全球发展提供了宝贵经验。 中国方面，政府通过发布一系列战略规划与政策文件，明确了自动驾驶的战略定位与发展方向。例 如，2025年8月，《关于深入实施"人工智能+"行动的意见》明确提出，"探索无人服务与人工服务相结 合的新模式"。11月，国务院办公厅《关于加快场景培育和开放推动新场景大规模应用的实施意见》， 全空间无人体系列为重点培育的22类新场景之一。同时，地方政府也积极响应，通过先行先试、政策创 新等手段，助力自动驾驶稳健发展。例如，北京、上海、武汉等城市纷纷建立了全链条的自动驾驶测试 与示范应用体系，为自动驾驶技术的落地提供了有力支撑。 近日，中国信通院2026深度观察报告会——人工智能法制专题报告会上，中国信通院政策与经济研 究所、人工智能与经济社会研究中心联合发布了《自动驾驶战略与政策观察（2025年）》专题报告。报 告核心观点指出，2025年全球自动驾驶产业正加速驶入发展快车道。这一趋势主要得益于两大驱动力： 一是感知、预测与端到端算法等核心技术的显著进步，有效提升了自 ...

【编者按】 站在"十五五"开局元年即将到来之时，我们关注如何推动中国式创新，发展新质生产力？破茧成蝶的路 上，中国企业家如何面对不确定性？又需要有怎样的精神内驱特质？ 自动驾驶出租车：安全吗？ 一场"无人"的马拉松：如何从0到1？ 这场突破背后是海量的复杂场景训练。小马智行在这场长跑中是已经积累了近6000万公里驾龄的"老司 机"。 回忆起公司发展历程中最激动的时刻，彭军的答案出人意料地具体到2022年底北京那个飘着小雪的日 子。那天启动无安全员的全无人运营测试，第一次实现了真正意义上"驾驶座空无一人"的突破。"有没 有安全员是0和1的本质区别。"彭军解释道。 T型战略布局：生根与拓展的双轨并行 无人驾驶出租车赛道上并不缺乏玩家。从萝卜快跑、文远知行到滴滴，从广汽、上汽集团到比亚迪，产 业链的四大领域：获客平台、AI司机、车辆生产和运营保养都会遇到对手和伙伴。小马智行该如何定 位自身角色？如何谋势布局？ 面对未来两三年规划，小马智行推出"T型战略"。纵向维度聚焦核心市场生根，提升服务密度。"如果路 上车太少，用户等半小时才叫到车，体验就无从谈起。"横向则指向更广阔的城市拓展，通过与地方政 府合作推动政策创新， ...

Core Insights - The Guangzhou Science and Technology Innovation and Entrepreneurship Competition (Guangzhou Double Innovation Competition) 2025 finals and INNO+ Greater Bay Area Science and Technology Carnival will be held on December 20 in Guangzhou, showcasing the achievements of the past decade and attracting over 50 investment institutions and 30 media outlets [1][4] - The competition has facilitated nearly 50 billion yuan in financing for participating companies over the past ten years, highlighting its role in transforming "dark horses" into market leaders [1] - Notable companies emerging from the competition include Yunzhou Biotech, valued at 7 billion yuan, and AiFeng GuangTong, which has made significant breakthroughs in 5G/6G communication filter technology [3][4] Event Structure - The event will feature three main components: final roadshow presentations, a decade of achievements exhibition, and special activities, emphasizing practical connections and ecological collaboration [5] - Approximately 300 guests are expected to attend, including representatives from national ministries, Hong Kong and Macau technology institutions, and financial organizations [5] - The event will include a pitch competition where nine winning companies will present their latest technological breakthroughs, followed by a voting session for the "Most Popular Award" [5] Special Activities - The event will host various "Science and Technology +" themed activities, such as roundtable discussions with entrepreneurs and scientists, international expansion initiatives, and deep learning sessions on the hard technology industry chain in Guangzhou [6] - The Guangzhou Science and Technology Bureau emphasizes that the competition serves as both a competitive platform and an accelerator for innovation [6]

【编者按】 全无人自动驾驶出租车已经在一些城市悄悄出现。小马智行就是其中一家，如今已率先在北上广深四大一线城市提供服务。继2024年在纳斯达克上市 后，小马智行近日在港交所上市，成为首批实现"美股+港股"双重上市的自动驾驶科技公司。无人驾驶安全吗？公众的接受度如何？创业公司在商业生态上 如何谋势布局？ 站在"十五五"开局元年即将到来之时，我们关注如何推动中国式创新，发展新质生产力？破茧成蝶的路上，中国企业家如何面对不确定性？又需要有怎 样的精神内驱特质？ 一场"无人"的马拉松：如何从0到1？ 这场突破背后是海量的复杂场景训练。小马智行在这场长跑中是已经积累了近6000万公里驾龄的"老司机"。 回忆起公司发展历程中最激动的时刻，彭军的答案出人意料地具体到2022年底北京那个飘着小雪的日子。那天启动无安全员的全无人运营测试，第一次实 现了真正意义上"驾驶座空无一人"的突破。"有没有安全员是0和1的本质区别。"彭军解释道。 自动驾驶出租车：安全吗？ 如果不是亲自体验，简直难以相信：无人的驾驶室，可以提速到最高时速80公里，遇到不走横道线横穿马路的人会提前停下，面对逆向行驶的车辆会及时 处理……各种丝滑的操作，让人既 ...

Group 1 - The core point of the article is the approval of the first batch of L3 autonomous driving models in China, indicating a significant step towards the practical application of autonomous driving technology in the country [2][4] - The approval process involved a thorough evaluation, including testing, expert reviews, and safety assessments, which signifies a shift from technical validation to mass production application in the autonomous driving industry [4][7] - The two approved models will be tested in limited areas and scenarios, and they cannot be used by individual car owners but will be operated by designated trial units, allowing consumers to experience autonomous driving through ride-hailing services [9][8] Group 2 - The legal framework surrounding autonomous driving needs to evolve, particularly in distinguishing responsibilities between human drivers and automated systems, especially in L3 to L5 levels of automation [12] - Future regulations should clarify the responsibilities in scenarios where human intervention is required, ensuring a clear distinction between the roles of the driver and the automated system [12] - The industry is moving towards a more refined responsibility allocation mechanism that aligns with the technological advancements in autonomous driving [12]

Core Insights - The approval of the first batch of L3-level conditional autonomous driving vehicle permits in China marks a significant milestone for the market, indicating a new phase for intelligent connected vehicles [2] - The rapid development of the autonomous driving industry is expected to positively impact the company's connected services and autonomous driving-related businesses [1][2] Group 1: Company Developments - The company has established itself as a leading player in the vehicle networking sector, leveraging technological innovations in areas such as dedicated vehicle networks, dual SIM connectivity solutions, and AI operations [1] - The company's connected vehicle business has shown rapid growth, with revenue contribution reaching 46% in the first half of the year [2] - The company’s subsidiary, Youka Technology, has achieved over 17 million connected vehicles, with more than 1.1 million connections overseas, indicating a strong international presence [2] Group 2: Industry Trends - The commercial rollout of L3-level autonomous driving is anticipated to significantly increase the demand for network stability, bandwidth, and low latency services [2] - The recent developments in the autonomous driving sector, including Tesla's Robotaxi testing in Texas, highlight the growing momentum in the industry [1] - The company is enhancing its global business strategy through the "Connection+" plan, collaborating with major global operators to create a unified global SIM solution [2]

Core Viewpoint - The article discusses the distinction between world models and end-to-end models in autonomous driving, clarifying that world models are not end-to-end but serve as a pathway to achieve end-to-end autonomous driving [2][3][4]. Group 1: Definitions and Concepts - End-to-end autonomous driving is defined as a model that processes information input on one end and outputs decision results without explicit information processing and decision logic [3]. - World models are defined as models that accept information input and internally establish a complete understanding of the environment, capable of reconstructing and predicting future changes [4]. Group 2: Course Introduction - A new course on world models has been launched, focusing on general world models, video generation, and OCC generation algorithms, including applications from Tesla and the Li Fei Fei team [5]. - The course aims to enhance understanding of end-to-end autonomous driving and is designed for individuals looking to enter the autonomous driving industry [15]. Group 3: Course Structure - Chapter 1 introduces world models and their relationship with end-to-end autonomous driving, covering historical development and current applications [10]. - Chapter 2 provides foundational knowledge on world models, including scene representation and relevant technologies like Transformer and BEV perception [10][16]. - Chapter 3 discusses general world models and popular algorithms such as Marble and Genie 3, explaining their core technologies and design philosophies [11]. - Chapter 4 focuses on video generation world models, detailing significant works and advancements in this area [12]. - Chapter 5 covers OCC generation models, discussing their applications and potential for trajectory planning [13]. - Chapter 6 shares industry insights and interview preparation tips for roles related to world models [14]. Group 4: Learning Outcomes - The course aims to elevate participants to the level of a world model autonomous driving algorithm engineer within approximately one year, covering key technologies and enabling practical application in projects [18].

Market Overview - The US stock market experienced a decline, with the Dow Jones falling by 0.47% to 47,885.97 points, the S&P 500 down by 1.16% to 6,721.43 points, and the Nasdaq decreasing by 1.81% to 22,693.32 points. Caterpillar and Nvidia led the decline, both dropping over 4% [5][11] - In the European market, the major indices showed mixed results, with the DAX down by 0.02% to 24,072.25 points, CAC40 down by 0.25% to 8,086.05 points, and the FTSE 100 up by 0.92% to 9,774.32 points [11] - The Asia-Pacific market saw the Hang Seng Index rise by 0.92% to 25,468.78 points, while the Nikkei 225 increased by 0.26% to 49,512.28 points [11] Economic Indicators - The UK’s November CPI growth rate slowed to 3.2%, the lowest since March, with a month-on-month decrease of 0.2% [11] - Japan's November exports to the US increased by 8.8%, marking the first rise in eight months, while exports to China fell by 2.4% [10][11] - China's fiscal revenue for the first 11 months reached 20.05 trillion yuan, a year-on-year increase of 0.8%, with tax revenue growing by 1.8% [12] Company Developments - BYD has initiated comprehensive testing for L3 level autonomous driving in Shenzhen, completing over 150,000 kilometers of real-world verification [17] - Xiaomi held its 2025 ecosystem partner conference, showcasing advancements in its open-source model, MiMo-V2-Flash, which ranks among the top in global open-source models [17] - CATL launched the world's first mass production line for humanoid intelligent robots in its facility, marking a significant milestone in intelligent manufacturing [17] A-Share Announcements - Xichuang Data plans to invest in the development and production of optical modules [17] - Shenzhen Urban Transport plans to raise up to 1.8 billion yuan for research and application of intelligent transportation equipment [17] - Shanghai Airport signed a contract for the transfer of duty-free store operation rights [17]

Group 1 - The core viewpoint is that Tesla's automotive business has been significantly downplayed, with its per-share value estimated at only about $30 to $40 [1] - Currently, Tesla's stock price reflects a bet on future potential that has not yet been fully realized, shifting focus from electric vehicle delivery volumes to advancements in robotics and AI [1] Group 2 - According to the analysis, autonomous driving now accounts for over 70% of Tesla's overall value, including the long-term commitment to the Robotaxi platform and the humanoid robot Optimus [3] - The growth of Tesla's energy business is notable, with its value now comparable to or possibly exceeding that of the automotive business [3] - When excluding the market's high expectations for autonomous driving software and energy storage, the remaining value of the electric vehicle business appears significantly overvalued relative to the current stock price [3] - The long-term energy strategy of Tesla is viewed positively, as the energy business is seen as a future profit engine that the market has not yet fully appreciated [3]

格隆汇12月18日｜中国L4级自动驾驶企业白犀牛宣布完成新一轮融资，本次由九坤创投、启赋资本、 元禾厚望、正景资本等共同注资。据白犀牛介绍，其在2025年内已完成3轮融资，全年累计融资额突破1 亿美元。 ...