1. 美国三大股指期货涨跌不一，纳指期货跌0.19%，标普500指数期货跌0.05%，道指期货涨0.08%。 2. 欧股主要指数普跌，德国DAX指数涨0.63%，英国富时100指数跌0.63%，法国CAC指数跌0.05%，欧 洲斯托克50指数跌0.19%。 17. 自动驾驶技术公司Mobileye将以9亿美元收购机器人公司Mentee Robotics。 美股时段值得关注的事件： 21:15美国12月ADP就业人数 23:00 美国12月ISM非制造业PMI/美国11月JOLTs职位空缺/美国10月工厂订单月率 (格隆汇) 3. 美国总统特朗普：委内瑞拉的临时管理当局将向美国移交3000万至5000万桶高品质、受制裁的石油。 4. 白宫：特朗普及其团队正在讨论多种取得格陵兰岛的方案，包括从丹麦购买该领土、与格陵兰建立自 由联系协定，军事手段亦被列为选项之一。 5. 美国高院将1月9日定为今年首个意见发布日，特朗普关税案有望迎来裁决。 6. 美联储理事米兰：2026年需大幅降息超100基点，直言现行政策正在抑制经济。 7. 黄仁勋回应H200何时售往中国：正加快生产，等待许可的最终细节敲定。 8. AMD苏姿丰 ...

Core Viewpoint - The article emphasizes the significant advancements in L4 and L3 autonomous driving technologies, highlighting a dual stronghold in the U.S. and China, with notable developments in commercial applications and regulatory approvals [3][10]. Group 1: December Smartization Marginal Changes Summary - L4 commercialization is experiencing global resonance, while L3 policies have made a "0-1" breakthrough. Tesla has initiated pure unmanned testing in Austin without safety drivers or passengers, and the Chinese version of FSD has completed compliance construction in Shanghai [3][10]. - Waymo's monthly active users have increased by 96.6% year-on-year to 1.64 million, accelerating the validation of its business model. Domestic players like Pony.ai have integrated into the WeChat ecosystem, and CAOCAO's "Ten Years, One Hundred Cities, One Hundred Billion" Robotaxi strategy continues to expand [3][10]. - The policy side for L3 has seen a milestone breakthrough with the issuance of the first batch of L3 permits by the Ministry of Industry and Information Technology, marking a transition from assisted driving to conditionally automated driving in China [3][10]. Group 2: January Smartization Catalysts - The industry may enter a year of Robotaxi explosion, driven by both technology and policy, accelerating towards an "iPhone moment" for AI smart vehicles. CES 2026 is expected to be a key industry indicator, with NVIDIA's CEO focusing on the latest advancements in physical AI [4][10]. - In the L4 dimension, attention is on Tesla's FSD V14.2 performance across the U.S. without driver intervention and the expansion of its fleet. For L3, January 2026 will mark the first complete month for public road commercial trials in cities like Chongqing and Beijing, with a focus on operational takeover rates and accident liability cases [4][10]. Group 3: Investment Recommendations - The company maintains a strong outlook for the L4 RoboX theme in 2026, favoring B-end software stocks over C-end hardware stocks. Preferred H-shares include Xpeng Motors, Horizon Robotics, Pony.ai, WeRide, CAOCAO, and Black Sesame Intelligence, while A-shares include Qianli Technology, Desay SV, and Jingwei Hirain [5][10]. - Relevant downstream application stocks include those in the Robotaxi sector, such as Tesla and Xpeng Motors, as well as technology providers like Horizon and Baidu. The transformation of ride-hailing and taxi services involves companies like Didi, CAOCAO, and others [5][10]. Group 4: Supply Chain Insights - In the upstream supply chain, key players include B-end unmanned vehicle manufacturers like BAIC BluePark, GAC Group, and Jiangling Motors, along with core suppliers for testing services, chips, domain controllers, sensors, and other components [5][10]. - The investment landscape is characterized by a spiral upward iteration of hardware and software, with significant opportunities in both sectors as the industry transitions from the 0-1 stage to more advanced phases [15][10].

Core Insights - Huang Renxun emphasized that robots are akin to AI immigrants, capable of taking on jobs that humans are reluctant to perform [1][5] - The company is considering restarting old production lines for GPUs, such as the RTX 3060, to incorporate new technologies due to current supply chain challenges [1][14][15] Group 1: AI and Robotics - The core theme of Huang's presentation at CES 2026 revolved around physical AI, including robotics and autonomous driving [5] - Huang stated that the "robot revolution" will drive economic growth, creating more jobs and absorbing more labor, while maintaining low inflation levels is crucial for a healthy economy [5] - He predicted that by the end of this year, robots will achieve human-level capabilities in mobility, joint movement, and fine motor skills [5] Group 2: Autonomous Driving - Huang introduced the world's first open-source, large-scale autonomous driving visual-language-action (VLA) inference model, named Alpamayo 1 [7] - The key distinction between Alpamayo and Tesla's FSD is that NVIDIA does not manufacture autonomous vehicles but provides a complete technology stack for companies developing them [9][11] - NVIDIA's systems have high industry penetration, with over 1 billion vehicles on the road today, and millions expected to have autonomous capabilities in the next decade [10] Group 3: GPU and Memory Challenges - NVIDIA did not release consumer-grade gaming graphics cards at CES 2026, citing rising memory costs and supply constraints as significant challenges [14][19] - Huang acknowledged the possibility of reviving production of older graphics cards to meet market demand, although this would require substantial R&D resources [15][16] - The company is positioned as a major buyer of high-bandwidth memory (HBM) and is collaborating closely with suppliers to ensure production aligns with product launches [22] Group 4: Future of Gaming - Huang discussed the future of gaming, suggesting that AI will play a significant role in character development and game realism, with advancements in rendering technology expected [19][20] - He believes that the gaming industry is entering an exceptional era, driven by AI innovations [20] Group 5: Leadership and Vision - Huang humorously shared two secrets for longevity as a CEO: not getting fired and avoiding boredom, while also acknowledging the significant responsibilities of leading NVIDIA [23]

Group 1 - The core viewpoint highlights the significant growth potential in the smart connected vehicle market, targeting a market size of 11.5 billion by 2030, with a critical period identified between 2025 and 2026 [1] - The solid-state battery sector is experiencing a global race for dominance, with key milestones set for equipment specification in 2026 and mass production in 2027, indicating three main investment themes [1] - A price increase of 60% by overseas giants has sparked a bullish trend in the storage chip market, confirming the ongoing prosperity cycle [1] - NVIDIA is intensifying its investment in autonomous driving, leading to a potential revaluation of the industry chain [1]

Core Insights - Mobileye has acquired Mentee Robotics for $900 million, which includes $612 million in cash and up to 26.2 million shares of Class A common stock, signaling a strategic move towards integrating humanoid robotics into its operations [4][6][8] - The acquisition is seen as a ticket to a new frontier, as Mobileye aims to combine autonomous driving with humanoid robotics, enhancing its capabilities in physical AI [3][18] - Mentee Robotics, founded in 2022, focuses on humanoid robot development and has a unique AI architecture that allows for efficient skill acquisition with minimal data, which could lower deployment barriers in various industrial settings [8][11][13] Mobileye's Strategic Direction - The acquisition is part of Mobileye's broader strategy to transition from a focus solely on autonomous driving to a more comprehensive approach to physical AI, which includes understanding context, reasoning intentions, and safe actions in the physical world [18] - Mobileye's core business remains strong, with recent contracts for its new generation of surround ADAS expected to deliver up to 9 million units [6][19] - The company plans to leverage its proven safety models and large-scale AI training infrastructure to support Mentee's humanoid robotics, while Mentee's technology could enhance Mobileye's autonomous driving capabilities [16][18] Future Developments - Mobileye's roadmap includes deploying Mentee's robots for concept validation in 2026, small-scale production in 2027, and large-scale deployment in structured environments by 2028, with plans to enter unstructured environments by 2030 [23][25] - The humanoid robots developed by Mentee are designed to be versatile, with capabilities suited for logistics and manufacturing, and are expected to play a significant role in the future of robotics [23][25] - The competitive landscape for Robotaxi services in the U.S. is set to intensify, with Mobileye planning to launch its services in collaboration with major partners, positioning itself against established players like Waymo and Tesla [21][25]

Group 1 - The U.S. House Energy and Commerce Committee plans to hold a hearing on January 13 to discuss legislation aimed at simplifying the deployment process for autonomous vehicles [1] - There has been a longstanding division in Congress regarding legislation to address barriers to the deployment of autonomous taxis [2] - The National Highway Traffic Safety Administration (NHTSA) currently maintains existing safety regulations and has not approved exemption requests from major automakers [2] Group 2 - The upcoming hearing will review multiple proposals, including one to increase the annual cap on the number of autonomous vehicles deployed to 90,000 [4] - Automakers have expressed concerns that certain safety standards, such as requiring rearview mirrors or steering wheels, may not be necessary for autonomous taxis [4] - One proposed bill would prohibit states from establishing rules regarding autonomous systems, while another would require NHTSA to develop guidelines for calibrating advanced driver assistance systems [4] Group 3 - The autonomous driving industry has faced increased scrutiny following an incident in October 2023, where a pedestrian was severely injured by a self-driving vehicle from General Motors [5] - Investigations have been launched by NHTSA into autonomous vehicles operated by Waymo and Amazon's Zoox [5] - Since the Trump administration took office, there has been a push to promote the adoption of autonomous driving technology in the U.S. [6] Group 4 - NHTSA has indicated it will expedite the review of automakers' applications to deploy autonomous vehicles without human control [7] - Tesla launched a small autonomous taxi service in Austin, Texas, equipped with a safety monitoring system [7] - Mercedes-Benz announced plans to introduce a new advanced driver assistance system in the U.S. later this year, allowing vehicles to operate autonomously in urban areas under driver supervision [7]

在最新的 CES 2026 上，英伟达（NVIDIA）首席执行官黄仁勋发表了主题演讲。这次发布会被业界视为从"生成式AI"转向" 物理AI "与" 推理型AI "的历 史性节点。 在本地发布会上，黄仁勋 宣布 Vera Rubin平台全面量产： 作为Blackwell 的继任者，Rubin架构正式亮相并宣布已进入全面投产阶段，预计2026年下半年 交付。其推理性能是前一代Blackwell的5倍，训练性能是Blackwell的3.5倍， 生成AI Token的成本直接降低至原来的1/10。 最关键的是，黄仁勋在演讲中强调，AI的下个阶段是物理AI，不仅要懂语言，也要懂物理世界，并发布了三款平台或应用模型： 1. Cosmos物理AI平台：英伟达发布了Cosmos系列基础模型（包括 Reason2、Predict2.5等）。它能通过视频和遥测数据学习物理规律，在虚拟的 Omniverse 空间中生成大规模合成数据。这意味着 AI 不再只是"猜下一个词"，而是能理解重力、碰撞和物体的物理属性。 2. Alpamayo推理型智驾模型：这是首个具备"思考"能力的自动驾驶模型。它不再仅仅依赖预设规则，而是能像人类一样对 ...

点击下方 卡片 ，关注" 自动驾驶之心 "公众号 戳我-> 领取 自动驾驶近30个 方向 学习 路线 >>自动驾驶前沿信息获取 → 自动驾驶之心知识星球 论文作者 | Xiaobao Wei等 编辑 | 自动驾驶之心 高斯泼溅的风，刮到了自驾的每个角落。 一大早看到了小米&杭电在泊车场景重建中的工作ParkGaussian。 相比英伟达3DGUT和OmniRe提升挺大，分享给大家。 泊车是自动驾驶系统（ADS）的关键任务，在车位拥挤且无GPS信号的环境中面临独特挑战。现有研究主要集中于二维车位感知、建图与定位，而三维重建领域的探 索仍显不足——该技术对于捕捉泊车场景中的复杂空间几何结构至关重要。单纯提升重建泊车场景的视觉质量并不能直接助力自动泊车，因为泊车系统的核心入口是 车位感知模块。 为解决这些局限，小米汽车联合杭州电子科技大学构建了首个专为泊车场景重建设计的基准数据集ParkRecon3D，其包含来自四台已完成外参标定的环视鱼眼相机的 传感器数据，以及密集的车位标注信息。在此基础上，本文提出了ParkGaussian框架，这是首个将3D高斯Splatting（3DGS）融入泊车场景重建的方案。为进一步提 ...

Group 1 - Hesai Technology has been selected by NVIDIA as a laser radar partner for the "NVIDIA DRIVE AGX Hyperion 10 platform," aimed at supporting L4-level autonomous driving [1][3] - The DRIVE Hyperion architecture is expanding its sensor ecosystem, covering cameras, millimeter-wave radar, laser radar, and ultrasonic technology, providing full-stack support for L4-level autonomous driving perception systems [3] - The DRIVE AGX Hyperion 10 platform features two NVIDIA DRIVE AGX Thor system-on-chip units, delivering over 2000 FP4 TFLOPS of computing power, sufficient for real-time data fusion from 360-degree sensors [3] Group 2 - NVIDIA's automotive VP emphasized that the DRIVE Hyperion platform is crucial for the transition to autonomous driving, integrating computing, sensors, and safety features into a unified open platform [4] - Hesai Technology will use its ETX laser radar as the core product for this collaboration, designed for L3 and L4 autonomous driving scenarios, featuring high resolution and long-range detection capabilities [4] - This partnership marks a significant milestone in the long-term collaboration between Hesai Technology and NVIDIA, reinforcing Hesai's position as a trusted core partner within NVIDIA's ecosystem [4] Group 3 - Looking ahead, Hesai Technology aims to collaborate with global automotive manufacturers to accelerate the marketization of next-generation autonomous driving technology, contributing to the upgrade of the global smart mobility industry [5]

Core Viewpoint - Huang Renxun emphasizes that robots are the "AI immigrants" capable of taking on jobs that humans are unwilling to do, highlighting the need for AI to support economic growth and job creation [10][11]. Group 1: AI and Robotics - Huang states that the "robot revolution" will drive economic progress and create more job opportunities while maintaining low inflation levels [11]. - He predicts that by the end of this year, robots will achieve human-level capabilities in mobility, joint movement, and fine motor skills [12]. - The development of robots requires not only visual perception but also tactile capabilities, which poses significant technical challenges [13]. Group 2: Autonomous Driving - Huang introduced the world's first open-source, large-scale autonomous driving visual-language-action (VLA) reasoning model, Alpamayo 1, and praised Tesla's FSD technology as world-class [15][16]. - NVIDIA's role is to provide a complete technology stack for companies developing autonomous vehicles, rather than manufacturing the vehicles themselves [16][20]. - The company has a high industry penetration rate, with over 1 billion vehicles on the road, and expects that millions will have strong autonomous driving capabilities in the next decade [20]. Group 3: AI Infrastructure and Memory Supply - Huang introduced NVIDIA's next-generation AI supercomputing platform, Vera Rubin, and discussed the challenges posed by rising memory prices and supply constraints [24][25]. - The company is positioned as a key player in the memory market, addressing the growing demand for high-bandwidth memory (HBM) and collaborating closely with suppliers to ensure production capacity aligns with product launches [36]. Group 4: Gaming and AI - NVIDIA upgraded its super-resolution model with the new DLSS 4.5 version, indicating a shift towards AI-driven gaming experiences [31]. - Huang predicts that future video games will be filled with AI characters, significantly enhancing realism and interactivity [32][33].