以"智驾赛道"为基本盘"，挖掘具身智能赛道新增长机遇，知行科技(01274)多条赛道纵深发展。 知行科技深耕于智驾赛道，是国内智能驾驶解决方案领头者之一，拥有已商业化的L2级至L2+级组合驾 驶辅助解决方案(L2+级并非SAE(1)标准下的官方分类)，并正在开发L3至L4级自动驾驶解决方案。凭借 着技术及产品力优势，该公司在智驾赛道上脱颖而出，落地场景从乘用车逐步渗透至商用车及物流车等 多领域。 智通财经APP了解到，近期该公司持续在物流车领域发力，2025年12月将自研的域控制器产品首次拓展 至无人物流小车领域，获得某头部品牌商用车项目定点，2026年1月再次获得商用车的量产项目定点， 为头部品牌无人物流配送小车提供基于地平线双征程6M(以下简称J6M)的自动驾驶软硬一体域控制器 iDC700。 域控制器为自动驾驶系统提供高效可靠的运行环境，是无人物流小车能够顺利从A点到B点的规模化基 石。该公司自研产品首次进入物流车领域且能够连续获得了汽车行业头部品牌客户订单，无人物流车领 域将成为公司智驾赛道新增长点。 知行科技主要通过与地平线合作模式切入无人物流车市场，此前在地平线J6M及征程系列平台上的开发 和量产 ...

Core Insights - The concept of "world model" has become a trendy term in the intelligent driving sector, with various companies like Xpeng, NIO, and Huawei adopting different terminologies for similar technologies [2][3][4] - World models are seen as a crucial component in the development of "physical world AI," enabling artificial intelligence to understand and replicate real-world dynamics [3][4] - The current application of world models in the intelligent driving industry is primarily cloud-based, with no direct implementation in vehicles yet [6] Group 1: Industry Trends - The shift from rule-based systems to AI-driven models in intelligent driving has led to a unified approach, where perception, prediction, and planning are integrated into a single network [7] - Despite the advancements, the transition to end-to-end models has revealed shortcomings in traditional simulation tools, necessitating the development of more sophisticated simulation environments [10][11] - The introduction of world models aims to address the limitations of existing simulators by providing a more comprehensive and realistic virtual environment for testing and validation [10][11] Group 2: Technical Challenges - The effectiveness of AI-driven models is hindered by the "black box" nature of end-to-end systems, making it difficult to diagnose errors and ensure reliability [9][10] - Current world models in the industry are still in the early stages, with limitations in generating realistic and diverse scenarios for training purposes [16][18] - The challenge lies in ensuring that generated scenarios accurately reflect real-world conditions, as inaccuracies can lead to poor model performance in practical applications [17][18] Group 3: Future Directions - Companies are exploring various approaches to enhance world models, with some opting for more controllable methods like 3D Gaussian reconstruction [14][15] - The ultimate goal is to develop world models that can support decision-making processes in vehicles, moving beyond their current use as training and validation tools [19] - Achieving a high level of accuracy and reliability in world models is essential for their deployment in real-world driving scenarios, which remains a significant hurdle for the industry [19]

Core Insights - The core focus of the articles revolves around the commercialization and strategic implications of L3 and L4 autonomous driving technologies in the automotive industry, highlighting the ongoing debate about the necessity and feasibility of L3 as a transitional phase towards full autonomy [1][2][3]. Group 1: Industry Trends and Developments - The penetration rate of L2 level assisted driving in China reached 64% with a year-on-year sales growth of 21.2% for new passenger cars equipped with such features in the first three quarters of 2025 [2]. - The Ministry of Industry and Information Technology (MIIT) of China issued the first batch of L3 conditional autonomous driving vehicle licenses in December 2025, sparking intense discussions about the commercial prospects of L3 technology [1][2]. - Companies like Xiaomi, XPeng, and GAC are actively pursuing L3 testing licenses, indicating a competitive race among automakers to establish their presence in the L3 segment [2]. Group 2: Strategic Perspectives - XPeng's CEO expressed that L3 is merely a transitional phase, suggesting that the industry may leap directly from L2 to L4, bypassing L3 altogether [3]. - The distinction between L3 and L4 is primarily based on the responsibility of the driver versus the vehicle, with L3 requiring the driver to take control under certain conditions, while L4 allows for full autonomy [3][5]. - The current regulatory standards for L3 certification are perceived as exceeding traditional definitions, necessitating that companies achieve near-L4 capabilities to obtain L3 licenses [5]. Group 3: Commercialization Challenges - The initial commercial value of L3 is challenged by restrictive operational scenarios, such as the limited use of the BAIC Arcfox Alpha S (L3 version) to specific road types and a maximum speed of 80 km/h, which may not attract average consumers [6]. - The integration of L2++ capabilities is seen as a prerequisite for achieving L4, indicating that the path to full autonomy may involve a gradual accumulation of data and experience in various driving scenarios [7]. - The automotive industry is experiencing rapid growth, with significant sales of vehicles equipped with advanced driving assistance features, which is fueling competition and innovation in the autonomous driving space [8].

中信里昂董事长李春波（左）和创业邦创始人兼CEO南立新（右）为周逵先生颁奖 1月15日，在第18届创业邦年会暨创业邦100未来独角兽大会上，红杉中国合伙人周逵获评为"2025 年度投资人"。 他在随后的演讲中分享了对创业者韧性与远见的观察。 在成功投出多家千亿市值公司的背后，周逵表 示，远见和胸怀、良好准备且行动 果断 、面对困难 积极和 有韧性，是这类成功者的典型特质。 于 投资者， 最好做个创业者背后的创业者，心有未来、 早 期 下注、从大的社会 需求 中去找大的商业 价值， 并协助 创业者 避坑和 聚焦。面对 今天 的成绩和运气， 企业家 保持 理性乐观和 "Day1"状 态，好事才会刚开始 …… 以下是周逵的演讲实录，由创业邦整理。 面对大目标，创业者要敢于早动手 摩尔线程、地平线、小马智行、一亩田 温暖祝福 最热的跑道会有最多人来支持，同时也有最多人来竞争。实现大机会的过程，没有不经历百般煎熬 的，需要founder具备非常的韧性和开阔的胸怀。最初你有远见和能力，别人包括投资人才会跟你 干，遇到困难你有韧性不放弃，同伴才不会放弃。你有远见且有胸怀，别人才愿意长期跟你一起打 拼。成功是founder有 ...

赵旭杨 开源证券汽车分析师: 好的，各位领导晚上好。我是开源证券汽车组的赵旭阳。今天感谢大家来参加我们今天的 智驾系列的第九期，关于特斯拉的一些智驾的一些信息的一些更新。以及就是近期那个， 特斯拉，就是美国的这个，高速公路安全管理局，其实也公布了一些智驾的数据。就是尤 其是这个 Robot Taxi 这边吧，一些这个，关于一些事故的一些数据的统计。这个之前其 实有一个这个比较大的一个新闻，然后大家可能觉得这个 Robot Taxi 的这个安全性可能 是产生了一些质疑。 然后这块我们也把这个数据下载下来，然后进行了一些这个分析。由于这个数据本身是包 含所有北美那边的这个 robot taxi 的车辆的，所以其实我们也可以整体性地观察出来的一 些，观察出来一些这个整体行业的一些这个情况。首先就是那个特斯拉的这个财报，财报 里边的话，其实我们觉得对于这个对于这个，我对这个少了一个环节，就是我们先还是把 我们的这个整体的支架的观点，简单给大家。汇报一下。我觉得，就是说整体智驾的这个 热度，我觉得这段这段时间是偏向于比较低位的一个状态的。 这个原因，我觉得主要还是说近期的这个智驾整体的这个催化呀，整体来说是比较小的。 ...

他是中国AI 1.0时代的拓荒者，在深度学习黎明前的黑夜，成为全中国第一个复现AlexNet （深度卷积神经网络模型） 、拿到AI"黄金榔头"的人。 他是 地平线创始团队成员及前智能驾驶总裁 ，却在公司2024年成功上市的高光时刻，转身闯进具身智能的无人区，去验证物理世界的Scaling Law。 他是 Vbot 维他动力 的创始人， 用60人团队 一年量产机器狗，首款 智能 四足机器人 产品发布即引爆，截止1月10日，预订订单量已经达到了6540台，并实现了 以C 端 带 动B端 的销量破局。 他是余轶南博士，Vbot 维他动力创始人兼CEO，他要打造全球领先的消费级机器人产品。 本期「善友探索流」播客，我们邀请到 Vbot 维他动力创始人兼CEO，同时也是混沌学园六期同学——余轶南博士。 在本场深度对谈中，余轶南博士贡献了极具 颠覆性的 三大商业洞见 ： 关于生存哲学 ：硬科技创业不是赌博，而是基于认知的精准判断。只有"锁死下限"，才敢去赌那个"无限的上限"。 关于公司本质： AI公司V.S.科技公司，区别不在技术，而在是否追求Scaling Law（尺度定律）带来的无限增长。 关于组织终局 ：他用了一个 ...

Sales Performance - In December, the total passenger car sales reached 2.317 million units, a month-on-month increase of 14.2% but a year-on-year decrease of 16.1% [4] - The new energy vehicle (NEV) market saw sales of approximately 1.33 million units, a year-on-year increase of 2.9% and a month-on-month increase of 9.2% [4] - NEVs accounted for 57.4% of total passenger car sales in December, a decrease of 2.6% from the previous month but an increase of 10.6% compared to the same month last year [4] New Energy Market Overview - In December, pure electric vehicle sales were approximately 778,000 units, a year-on-year increase of 0.9% and a month-on-month increase of 1.1% [9] - Plug-in hybrid sales reached about 552,000 units, with a year-on-year increase of 5.9% and a month-on-month increase of 23.0% [9] - Cumulative NEV sales for 2025 reached 12.372 million units, with a cumulative year-on-year increase of 15.1% [9] Top Cities for NEV Sales - The top 10 cities accounted for 29.6% of NEV sales, an increase of 2.4% from the previous month [10] - The top three cities for NEV sales were Shanghai, Hangzhou, and Guangzhou, with Dongguan and Tianjin replacing Wuhan and Xi'an in the top ten [10] - The highest NEV penetration rates were in Shenzhen (71.0%), Shanghai (68.8%), and Tianjin (68.1%), while Dongguan had the lowest at 53.8% [10] Pure Electric Market Analysis - The top three segments in the pure electric market in December were B-SUV (19.5%), C-SUV (15.6%), and A0 class (14.3%) [14] - The market shares for B-SUV and C-SUV increased significantly by 3.7% and 2.5% respectively, while the A00 class market share decreased by 8.2% [14] - Compared to the same month last year, the C-SUV market share expanded from 7.2% to 15.8%, while the A00 class market share shrank from 13.8% to 5.2% [14] Market Dynamics - Key industry events in December included the global launch of the Xiaopeng X9 super range extender and the introduction of Huawei's next-generation DriveONE range extender generator [21] - GAC Group's internal restructuring led to the integration of its two NEV brands, Haobo and Aion, into a new business unit [23] - SAIC's semi-solid-state battery achieved mass production and delivery, marking it as the world's first and only semi-solid-state battery to be mass-produced [25] Industry Policy - The Ministry of Industry and Information Technology released a digital transformation implementation plan for the automotive industry on December 30, aiming to promote digitalization and intelligent upgrades [28] - The plan includes six key tasks, such as establishing a diagnostic assessment system and promoting the application of artificial intelligence in the automotive sector [29] - By 2027, the plan aims to improve labor productivity in the industry by 10% and shorten product development and delivery cycles by 20% [29]

Core Viewpoint - The Vice President of Horizon, Lv Peng, affirmed that the company will not abandon its end-to-end team even if it launches the VLA, emphasizing that a strong end-to-end foundation is essential for the success of VLA [1] Group 1 - Horizon's commitment to maintaining its end-to-end team is seen as crucial for the development of its new VLA product [1] - Lv Peng believes that without a solid end-to-end system, the VLA would struggle to perform effectively [1]

Core Viewpoint - The interview highlights the current divergence in smart driving technology routes, emphasizing that there is no need for terminology anxiety as various approaches like end-to-end, VLA, WA, and VA are fundamentally aligned in their technical architecture [1] Group 1: Technology Perspectives - The market should not be concerned about the different terminologies in smart driving technology, as they are all fundamentally based on an end-to-end architecture [1] - End-to-end is considered the cornerstone for integrating new modalities and enhancing product performance, indicating its critical role in the development of smart driving technologies [1] - If end-to-end is not executed well, it will hinder the effectiveness of VLA, suggesting a strong interdependence among these technological approaches [1]

如何分辨一段式和两段式端到端？地平线副总裁吕鹏称，消费者不太需要关注这些名词。对于消费者角 度来说，最直观的就是开车时他是否真的愿意在城市里去使用这个系统，有没有足够的安心？对于从业 者来说，分辨起来其实没有那么复杂，主要看模型输出轨迹的连贯性和拟人性好不好，特别是交互情况 下模型的整个动作是按照场景step by step的，还是行云流水的。这是一段式端到端中很重要的判断点。 只有行云流水的时候，里面才没有太多的规则、行动比较一致，不会有场景的割裂感。 (作者：易思琳编辑：张明艳) ...