三位产业领军人物，揭秘中国制造的智能化进阶之路。 从吉利汽车研究院常务副院长任向飞，到小马智行创始人兼首席技术官楼天城，再到中坚 科技副总经理兼首席战略官鲍嘉龙，在2025世界人工智能大会的特别节目《智能驱动新经济： 从自动驾驶到具身智能的产业新纪元》中，他们展开了一场关于"AI重塑中国制造 " 的深度对 话。 当汽车不再只是出行工具，当机器人开始理解物理世界，一场万亿级的产业革命正在中国 加速展开。 这不仅是传统制造业的数字化转型，更是中国智造迈向全球价值链顶端的关键跳跃。 传统车企的"三驾马车 " 算力、算法、数据全面赋能 " 智能汽车的发展方向就是从普通的出行工具要向具身智能体的形态去转变。" 吉利汽车研究院常务副院长任向飞开门见山地指出了汽车产业的未来走向。 在这场对话中，传统车企不再是简单的制造商，而是AI技术的深度整合者。任向飞透露，吉利在2021 年就开始部署星睿智算，目前拥有23.5个EFlops的算力水平，在汽车行业中位居前列。 从动力域到智能驾驶，从娱乐域到底盘域，全域AI正在重新定义汽车这个复杂工业品的每一个环节。 传统车企如何通过AI实现从制造到智造的华丽转身。 自动驾驶商业化拐点 成 ...

Group 1: Macro and Financial Data Insights - In July, the growth of M1 and M2 exceeded market expectations, with M2 expanding by 8.8% year-on-year and M1 growing by 5.6%, up from 8.3% and 4.6% in June respectively [2][3] - New social financing in July was 1.16 trillion yuan, lower than the Bloomberg consensus of 1.63 trillion yuan, while new RMB loans decreased by 500 million yuan, indicating a shift in financing structure and seasonal factors [2][3] - The stock of social financing grew at a rate of 9.0% year-on-year, an increase from 8.9% in June, with seasonally adjusted month-on-month growth rising from 8.4% to 9.6% [2][3] Group 2: Banking Sector Analysis - The July social financing increment of 1.16 trillion yuan was below the expected 1.41 trillion yuan, with a year-on-year increase of 389.3 billion yuan [5] - The government bonds were the main support for social financing in July, while M1 growth showed a marginal recovery [5] - A new consumption loan subsidy policy is expected to stimulate the growth of consumer loans, indicating a positive outlook for the banking sector [5] Group 3: Company-Specific Insights - Tencent's Q2 revenue grew by 14.5% year-on-year, exceeding consensus expectations, with significant growth in value-added services, advertising, and fintech revenues [11] - The company is expected to benefit from the upcoming launch of several major shooting games, which could drive both player engagement and monetization [11] - Huatai Securities initiated coverage on Yuntianhua with a "buy" rating, citing its leading position in the phosphate industry and expected steady demand growth for fertilizers [15] Group 4: Technology and Robotics - The introduction of teaching-free robots is transforming the welding industry, addressing labor shortages and improving efficiency through advanced visual systems and welding software [7] - These robots are expected to penetrate more complex applications, such as shipbuilding, as technology continues to evolve [7] Group 5: Consumer and E-commerce Trends - SEA's Q2 revenue reached $5.26 billion, a 38.2% year-on-year increase, driven by strong performance in e-commerce and digital financial services [29] - The company anticipates continued growth in its e-commerce GMV, projecting a 25% year-on-year increase for Q3 [29] - Tencent Music's Q2 revenue was 8.44 billion yuan, up 17.9% year-on-year, benefiting from rapid growth in super memberships and strong performance in non-subscription services [27]

除了热情的观众，在一级市场，具身智能的投资更是一浪更比一浪强。 今年7月，京东一天官宣投资三家机器人公司，分别为众擎机器人、千寻智能和逐际动力。更早一个 月，银河通用宣布完成11亿元新一轮融资，据第三方机构统计，这是迄今为止国内具身智能领域单笔最 大融资。 更多的公开数据则显示，2025年上半年在具身智能的融资规模，已经超过了去年全年。截至8月12日， 国内现存机器人相关企业95.8万家。今年前7月已注册机器人相关企业15.28万家，较去年同期相比，增 长43.81%，增速远超去年全年注册水平。 而随着资本进场和大批机器人相关企业入局，机器人这个全新的行业生态正在加速形成，并正式开启了 上半场的征程。 热钱进场，资本先行站位 判断一个早期行业是否具有足够大的商业前景时，资本永远是最具参考意义的风向标。 一场世界机器人大会，彻底引爆了普通大众对于人形机器人的关注度。 据世界机器人大会的官方数据统计，本次展会的参观人数累计突破了130万。据笔者现场看到的情况， 几乎每一个展台，都被热情的观众给层层叠叠的包围住。尤其是在机器人拳击赛、足球赛的展示环节， 众多展台更是人满为患。 高工机器人产业研究所数据显示，国内具身 ...

Investment Rating - The investment rating for the company is "Buy-A" (maintained) [1] Core Views - The company reported a revenue of 597 million yuan for the first half of 2025, representing a year-on-year growth of 10.32%. However, it incurred a net loss attributable to shareholders of 14.19 million yuan, an improvement from a loss of 26.24 million yuan in the same period last year [4] - The company's performance in the smart driving solutions, consumer electronics, and PCB sectors showed significant growth, with visual measurement equipment revenue increasing by 72.44% year-on-year [4][5] - The semiconductor sector has made significant breakthroughs, with the company’s joint venture achieving major progress in the field [5] Financial Performance - The company’s revenue for Q2 2025 was 378 million yuan, a year-on-year increase of 8.76%, with a net profit attributable to the parent company of 18 million yuan, up 54.07% year-on-year [4] - The company’s net profit forecast for 2025-2027 is expected to be 170 million yuan, 180 million yuan, and 200 million yuan respectively, with year-on-year growth rates of 32.8%, 11.0%, and 10.0% [7] - The earnings per share (EPS) for 2025 is projected to be 0.86 yuan, with a price-to-earnings (P/E) ratio of 60.8 based on the closing price of 52.00 yuan on August 13 [7] Business Segments - The company’s visual measurement equipment generated revenue of 226 million yuan, driven by growth in major consumer electronics clients [4] - The smart driving solutions segment saw a revenue increase of 125.39% year-on-year, attributed to growth in automotive smart driving and humanoid robot sectors [4][5] - The visual detection equipment revenue decreased by 70.81% due to a decline in the photovoltaic industry [4] Market Position - The company is recognized as a leading visual equipment platform in China, focusing on the integration of artificial intelligence technology to drive industrial digital transformation [6] - The company is expanding its business collaborations in the smart driving sector, particularly with major automotive manufacturers [5][6]

（文/孙梅欣 编辑/张广凯） 尽管搭上了"机器人"概念的东风，但最近10年都没盈利的天链机器人，冲击上交所被"劝退"的风险大大 增加。 虽然天链机器人不属于科创板负面清单行业，但长期无法盈利，可能是导致企业难登上交所的因素。 公告显示，2023年-2024年期间，天链机器人营收分别为2139.26万元、2991万元，经营活动产生的现金 流量净额分别为-1.44万元、-908.84万元。同时，归属于挂牌公司股东的净利润均为负值，营业收入最 近三年复合增长率为12.03%。 公告称，截至2025年7月30日，公司总市值为7.56亿元，公司的市值及财务等指标暂不满足《上海证券 交易所科创板股票上市规则》、《科创属性评价指引(试行)》规定的上市条件。 根据相关规则，在上交所科创板上市的股票，至少要符合多项标准中的一项。其中，包括预计公司市值 不低于人民币10亿元，要求最近两年净利润均为正，且累计净利润不低于人民币5000万元等条件。 另外根据《试行》要求，在科创板上市企业，最近三年营业收入复合增长率要达到25%，或最近一年营 业收入金额达到3亿元。 由此来看，天链机器人在市值、净利润、复合增长率等多个方面的财务指标上 ...

Core Insights - The World Robot Conference (WRC) showcased major manufacturers and suppliers in the robotics industry, with a significant focus on LiDAR technology as a standard for embodied intelligence [1] - Hesai Technology anticipates that its LiDAR shipments in the robotics sector will exceed 200,000 units by 2025, surpassing the total historical deliveries of robotic LiDAR [1] - The demand for robotic lawn mowers is surging, with a notable increase in orders expected this year [1] Industry Trends - The rise of boundary-less robotic lawn mowers has led to a rapid increase in domestic players entering the market, with 32 companies having entered the sector over the past five years [2] - Current global sales of robotic lawn mowers are approximately 1.8 million units, with a penetration rate of less than 10%, indicating the industry is still in its early stages [2] - The combination of hardware and software optimization is crucial for the success of robotic lawn mowers, particularly in North America where grass quality and growth rates demand high-performance equipment [2] Company Developments - Hesai Technology signed a strategic cooperation agreement with Keting Technology to provide 300,000 JT series LiDAR units for their robotic lawn mowers within a year [3] - The JT series LiDAR has achieved mass production of 100,000 units in less than five months, primarily for the lawn mower sector [3] - Suton Juchuang has completed a full-link solution layout involving hardware, chips, and AI technology, with expectations to reach a shipment volume of 100,000 units by 2025 [3] Product Innovations - At the WRC, Hesai Technology showcased two core product solutions: the mini ultra-spherical 3D LiDAR JT series and the next-generation automotive-grade solid-state LiDAR FTX [4] - Suton Juchuang introduced the Active Camera platform, with the second product AC2 set to launch in the second half of the year [4] - The demand for LiDAR is increasing as it is considered an essential sensor for robotic mobility due to its wide sensing range, compact size, and reliability [5] Market Outlook - The demand for embodied intelligence in robotics is expected to grow, with various manufacturers planning to launch humanoid robots by the end of 2024, marking 2025 as a potential commercialization year for humanoid robots [6] - The market for robotic lawn mowers alone could reach millions of units, with the overall market size for service robots projected to surpass that of the automotive sector [6]

行业资讯： 相关ETF：公开信息显示，科创半导体ETF跟踪上证科创板半导体材料设备主题指数，囊括科创板中半 导体设备（59%）和半导体材料（25%）细分领域的硬科技公司。半导体设备和材料行业是重要的国产 替代领域，具备国产化率较低、国产替代天花板较高属性，受益于人工智能革命下的半导体需求扩张、 科技重组并购浪潮、光刻机技术进展。 1.正帆科技公告称，公司已与辽宁汉京半导体材料有限公司5名股东签署《股份转让协议》，购买上述5 名股东持有的汉京半导体总计62.2318%的股权，转让价格合计为11.2亿元。本次交易高度契合公司发展 战略，进一步推动公司OPEX业务的发展，交易完成后，汉京半导体将作为公司的控股子公司，纳入公 司合并报表范围。 半导体材料ETF及其联接基金（A类：020356、C类：，指数中半导体设备（59%）、半导体材料 （24%）占比靠前，充分聚焦半导体上游。 2.艾为电子2025年上半年实现营业收入13.70亿元，同比下降13.40%；归母净利润1.57亿元，同比增长 71.09%；扣非净利润1.23亿元，同比增长81.88%。艾为电子表示，利润增长主要得益于公司通过对高性 能数模混合、电源管理 ...

温馨提示 ： 点击下方图片，查看运营团队最新原创报告（共235页） 说明： 欢迎约稿、刊例合作、行业交流 ， 行业交流记得先加入 " 机器人头条"知识星球 ，后添加（ 微信号：lietou100w ）微信； 若有 侵权、改稿请联系编辑运营（微信：li_sir_2020）； 星海图G0： 端到端双系统全身智能 VLA模型，以"慢思考、快执行"双系统为架构基础，叠加500小时高一致性全身动作 数据和"三阶段学习训练"策略，实现从视觉和语言指令到23自由度全身控制的长程任务执行。 G0大模型 采用"双系统"架构， 将机器人控制分为高层思考与低层动作两个模块；高层的 G0-VLM是一个视觉语言模 型，负责语言理解、 复杂逻辑推理 和任务拆解、规划， 以2 Hz的频率输出子任务指令； 低层的 G0-VLA 则是视觉-语 言-动作模型，负责感知融合和 全身控制动作与移动轨迹， 动作执行频率 15 Hz 、运动频率200Hz。 —— 两者异步运行：从而实现全局规划的智能性与局部执行的高响应性。 据官方表示： 全身动作、长续任务、柔性操作、语言理解方面表现，G0的评测数据已压过 π0。 正文： 融资超15亿元，估值10亿美元 ...

点击下方 卡片 ，关注" 具身智能 之心 "公众号 作者丨 Yongjie Bai等 编辑丨具身智能之心 本文只做学术分享，如有侵权，联系删文 >> 点击进入→ 具身智能之心 技术交流群 更多干货，欢迎加入国内首个具身智能全栈学习社区 ： 具身智能之心知识星球 (戳我) ， 这里包含所有你想要的。 研究背景与动机 现有视觉-语言-动作（VLA）模型在多任务机器人操作中，常依赖固定视角和共享视觉编码器，这会限制3D感知能力并引发任务干扰，影响鲁棒性和泛化性。 固定视角在复杂场景中问题突出：在杂乱或动态环境中，固定相机易导致目标物体或末端执行器被遮挡，使场景理解不完整，进而导致动作预测不准确。例 如"把糖放进橱柜"的任务中，固定摄像头的前视图可能只拍到橱柜，左右肩视图只拍到被夹持的糖，无法同时覆盖关键信息，可能导致操作失败。 共享编码器的局限同样明显：即使是语言条件的 Transformer 方法（如 RVT、RVT-2），其共享编码器在处理视觉和语义差异大的任务（如"摘苹果"与"开抽屉"） 时，会因任务干扰限制模型在多任务中的泛化性和扩展性。 核心方法：TAVP 框架 为解决上述问题，我们提出任务感知视图规划（ ...

Core Viewpoint - Nvidia is significantly advancing its robotics development infrastructure, focusing on the integration of AI and computer graphics to enhance robotic capabilities and reduce training costs [17][20][21]. Group 1: Product and Technology Updates - Nvidia introduced the upgraded Cosmos world model at the SIGGRAPH conference, which is designed to generate synthetic data that adheres to real-world physics [2][3]. - The upgrade emphasizes planning capabilities and generation speed, with enhancements across software and hardware, including the new Omniverse library and RTX PRO Blackwell servers [4][8]. - The new Cosmos Reason model features 70 billion parameters and reasoning capabilities, aiding robots in task planning [6][10]. - Cosmos Transfer-2 and its lightweight version accelerate the conversion of virtual scenes into training data, significantly reducing the time required for this process [12][13]. Group 2: Integration of AI and Graphics - Nvidia's AI research vice president highlighted the powerful synergy between simulation capabilities and AI system development, which is rare in the industry [5]. - The combination of Cosmos and Omniverse aims to create a realistic and scalable "virtual parallel universe" for robots to safely experiment and evolve [22][23]. - The integration of real-time rendering, computer vision, and physical simulation is essential for building this virtual environment [23]. Group 3: Market Strategy and Collaborations - Nvidia is strategically positioning itself in the robotics sector, recognizing the trend of merging computer graphics with AI as a transformative force in the industry [20][21]. - The company is collaborating with various Chinese firms, including Alibaba Cloud and several robotics companies, to expand its influence in the domestic market [26][27]. - Nvidia's approach mirrors its previous strategies, where it provided computational resources to emerging AI companies, indicating a similar trajectory in the robotics field [25][26].