Core Insights - The first humanoid robot half marathon took place in Beijing, featuring 20 teams of humanoid robots competing alongside running enthusiasts, with the "TianGong Ultra" robot winning the race in 2 hours, 40 minutes, and 42 seconds [1][6]. Event Overview - The half marathon course spanned 21.0975 kilometers, starting at Nanhai Park and ending at Tongming Lake Information City, incorporating diverse terrain including flat areas and slopes with a maximum incline of 9 degrees [2]. - The event showcased 18 different robot models, with notable entries from Shenzhen-based companies, including the champion "TianGong Ultra" and other robots from companies like Zhongqing Robotics and Leju Robotics [2][3]. Technological Innovations - The "KuaFu" robot, a versatile humanoid capable of walking, jumping, and performing tasks, utilized advanced algorithms combining model predictive control and imitation learning to enhance its movement efficiency and stability [3]. - "TianGong Ultra" featured high-efficiency joint control and a lightweight design, enabling it to maintain balance and speed during the race, achieving a maximum speed of 12 kilometers per hour [5][6]. Industry Implications - The event marked a significant step in the practical application of humanoid robots, transitioning from laboratory settings to real-world environments, providing valuable data on stability, endurance, and path planning [7]. - The competition highlighted the ongoing advancements in humanoid robotics, with major companies like Tesla, Figure, and UBTECH releasing new generation products, indicating a shift from concept validation to real-world applications [7][8]. - The event served as a demonstration for standardized testing scenarios, promoting the integration of cutting-edge technology and its potential applications in various fields such as security, home services, and industrial collaboration [8].

用不同的力度来测试机器人的每一根手指，可以看出小拇指的反馈力度会更强烈一些，这样灵敏的动作背后是因为它指尖的这个地方集成了密 密麻麻的柔性传感器，这么小的面积最多可以集成几百个柔性传感器。 这是一个传感器的放大版本，它非常轻、非常薄，厚度只有0.2毫米左右，跟我们人类皮肤表层的厚度差不多，而且它非常柔软，可以随意弯 曲。如果把这样的柔性传感器和经过特殊工艺处理的膜布两者相结合，就可以做成"电子皮肤"。 央视网消息：要想让机器人更快走入工厂和家庭，完成接近人类的各种复杂动作，触觉感知必不可少。近期，不少人形机器人披上了"电子皮 肤"，加快对真实世界的探索。 装上"电子皮肤"和仿真硅胶表层后，机器人手指的触感和灵敏度更接近人类。除此之外，电子皮肤还可以用在机器人身体的不同位置，感知温 度和距离。 某企业首席科学家陈书厅介绍，"电子皮肤"的需求正在指数级增长阶段，在几年之前可能是一个相对小众的产品，现在需求非常多，预计近5 年它的年增长率应该会增加50%以上。 "电子皮肤"订单爆发式增长 企业扩产 价格下降 从小众产品到机器人的刚需，"电子皮肤"应用快速增长，生产研发也在提速。据了解，目前，我国在"电子皮肤"领域具 ...

Group 1 - The liquidity of the Robot Index ETF showed a turnover of 3.12% with a transaction volume of 18.34 million yuan, indicating active trading [2] - The Robot Index ETF experienced a significant growth in scale, increasing by 82.70 million yuan over the past two weeks, and its shares grew by 54 million, reflecting strong investor interest [2] - The Robot Index ETF closely tracks the CSI Robot Index, which includes companies involved in system solutions, digital workshops, automation equipment manufacturing, and other robot-related sectors [2] Group 2 - The latest report from Zhongtai Securities highlights that "embodied intelligence" has been included in the government work report, indicating a strategic emphasis on this area [3] - The robot industry is expected to see a new wave of growth driven by the mass production of Tesla's Optimus in 2025, marking a year of performance realization for companies in the robot supply chain [3] - Domestic automotive companies are actively engaging in both self-research and collaborative models, which are likely to be implemented in production scenarios [3]

Core Viewpoint - The discussion surrounding the Yushu H1 robot's ability to perform complex gymnastic movements while struggling in a half marathon highlights the "impossible triangle" of hardware, algorithms, and scenarios in the humanoid robotics field [2] Group 1: Hardware Design Challenges - The Yushu H1 features a joint motor torque density of 230Nm/kg and a real-time control system with millisecond response speed, allowing it to switch its center of gravity in 0.5 seconds, prioritizing instantaneous power output [3] - However, this design leads to a significant power consumption of 300W per joint during dynamic activities, which is much higher than during normal walking [3] - During a marathon at a speed of 6km/h, the robot requires over 500W of total heat dissipation, exceeding its passive cooling system's capacity, resulting in knee joint temperatures exceeding 80°C after one hour, causing a 23% decrease in torque precision [4] Group 2: Algorithmic Layers - The VSLAM (Visual Simultaneous Localization and Mapping) system developed by Yushu allows for environmental perception at 60 frames per second, with a gait generator trained through reinforcement learning, achieving a reaction delay of only 80ms when avoiding obstacles, close to human spinal reflex speeds [5] - For marathon running, an energy consumption model must be established, requiring the robot to keep energy use below 238Wh per kilometer to cover 21 kilometers with a total battery capacity of 5kWh [7] - If the path planning algorithm is not optimized, actual endurance may drop to 15 kilometers, and the robot's "5-minute battery swap" plan results in loss of motion memory, effectively resetting its learning model every 5 kilometers [7] Group 3: Differences in Participants - The Yushu H1 robot participating in the marathon was modified by a client and did not use the original factory algorithm, which is based on 100,000 hours of simulation training with a foot pressure sensor error compensation frequency of 200Hz [8] - In contrast, the third-party algorithm used had a sensor calibration frequency of only 50Hz, leading to a fourfold increase in error accumulation during continuous running, explaining why the robot could stand on one leg for over 30 minutes in the lab but struggled in the race [8] Group 4: Industry Insights - The trend towards performance specialization is evident, with Tian Gong Ultra reducing weight by 8kg and lowering energy consumption by 15% for inspection and logistics, while Yushu H1 retains a 12-degree-of-freedom dexterous hand for flexibility in rescue and service scenarios [9] - Technological breakthroughs are being pursued, such as Boston Dynamics' Atlas experimenting with liquid metal cooling systems to reduce joint operating temperatures by 40%, and Tesla's Optimus improving energy efficiency through silicon carbide inverters [9] - Over the next five years, humanoid robots are expected to achieve breakthroughs in both gymnastics and marathon capabilities as solid-state battery energy density surpasses 400Wh/kg [9] Conclusion - The evolution of technology is a necessary path, as seen in the transition from laboratory demonstrations to practical applications in humanoid robotics, emphasizing the importance of algorithm optimization for performance leaps [10]

Group 1 - The Beijing Yizhuang Half Marathon and Humanoid Robot Half Marathon took place on April 19, with the "Tiangong" robot from Beijing winning the competition [1] - The latest public fund quarterly report indicates that "robots" have become a high-frequency keyword, with multiple robot concept stocks being increased by funds in Q1, contributing significant returns [1] - As of April 21, 2025, the National Securities Robot Industry Index (980022) rose by 1.22%, with component stocks such as Guirui Technology (600562) up by 10.01% and Aobi Zhongguang (688322) up by 6.71% [1] Group 2 - As of March 31, 2025, the top ten weighted stocks in the National Securities Robot Industry Index (980022) accounted for 42.06% of the index, including companies like Huichuan Technology (300124) and Keda Xunfei (002230) [2] - The Ministry of Industry and Information Technology reported that nearly 70% of key products in the equipment industry saw production growth in Q1, with a growth rate of 69.2%, an increase of 16.7 percentage points from the previous year [2] Group 3 - CITIC Securities research report highlights the growing attention on humanoid robots, suggesting that humanoid robots are the best carriers of embodied intelligence and are central to the US-China AI technology competition, indicating a broad long-term potential [3] - The report recommends exploring investment opportunities in "AI + robots" beyond just humanoid robots, focusing on faster commercialization in areas like exoskeleton robots, dexterous hands, and sensors [3]

Core Insights - The Chinese robotics industry is projected to exceed a market size of 155 billion yuan in 2024, following a market size of over 130 billion yuan in 2023, with a compound annual growth rate (CAGR) of 19.46% over the past five years [1][3]. Market Overview - The robotics market in China is categorized into industrial robots, service robots, and special robots, with service robots holding the largest market share at approximately 45% in 2023 [3]. Industry Competition - The competition in the Chinese robotics industry is intensifying, featuring major players such as Shenyang Machine Tool, Gree, Ecovacs, and DJI, who are innovating in automation, intelligence, and efficiency to enhance their market competitiveness [6][7]. Key Companies - **Siasun Robot & Automation**: A leading provider of industrial robots and automation solutions, serving various sectors including automotive and electronics [7]. - **Gree Electric Appliances**: A major player in the home appliance industry, increasing its focus on industrial robots and smart manufacturing through acquisitions [7]. - **Shenyang Machine Tool**: An established company in the machinery manufacturing sector, specializing in CNC machine tools and industrial automation technology [7]. - **Ecovacs Robotics**: Focused on smart home and cleaning robots, offering products that enhance household automation and quality of life [7]. - **DJI Innovations**: A global leader in drone and robotics technology, developing products for agriculture and security, with an emphasis on precision agriculture and automated inspections [7].

每经记者 张韵 叶晓丹 每经编辑 马子卿 3月末，苏州的空气中还带着些许凉意。步入尚金湾总部科技园的大门，络绎不绝的参访者热情如潮，围绕具身智能应用落地的交流声此起彼伏。从远处跑 来的机器人"小麦"手指握拳、上下摆臂、腰板挺直，以每秒2米的移动速度，享受着众人的围观。 "2023年以来，人工智能的应用能力得到巨大发展。今年，机器人本体的运动能力出现巨大提升。"魔法原子副总裁兼研发负责人陈春玉说，当具身智能技术 范式发生巨变，端到端的技术路线让机器人拥有了自主能力，使得应用落地加快释放，这超出了业内预期。 小麦是魔法原子"一手带大"的人形机器人。别看公司成立于2024年，早在数年前其创始团队就开始涉足机器人研发。如今，大模型技术的发展为人形机器人 带来了更多的市场前景，也给魔法原子在工业、商业等场景落地提供了广阔的想象空间。 尽管现阶段具身智能概念火爆异常，商业化路径仍面临诸多争议，机器人走进家庭生活究竟还有多远？每日经济新闻《机器人产业链一线调研》第二期实地 探访魔法原子，近距离感受具身智能在未来五年的技术迭变与更多可能性。 本体轻量化趋势 "今年以来，跑酷、空翻等高难度动作集中涌现，人形机器人正在用颠覆性的 ...

Market Overview - The AI and robotics sector experienced a sideways adjustment, with the Huaxia Sci-Tech AI ETF (589010) declining by 0.9% as of April 18. Key stocks such as Chipone Technology and Anlu Technology led the decline [1] - The Robotics ETF (562500) remained flat, with a trading volume of 415 million yuan and a turnover rate of 3.57%, indicating active market trading [1] Key Events - On April 19, the world's first humanoid robot half marathon was successfully held in Beijing, with 20 robot teams participating. The TianGong Ultra robot from the Beijing Humanoid Robot Innovation Center finished first, completing the 21-kilometer race in 2 hours and 40 minutes [1] - The CTO of the Beijing Humanoid Robot Innovation Center, Tang Jian, stated that the price of humanoid robots is expected to be comparable to that of an entry-level sedan as they become more widely adopted [1] Institutional Insights - Dongwu Securities highlighted several issues with humanoid robot hardware revealed during the marathon: 1. Insufficient battery life, as most robots required battery changes mid-race except for the walking robot [3] 2. Inadequate joint cooling, with robots needing cooling sprays to maintain joint functionality during the competition [3] - The event, hosted by local government and broadcasted by CCTV, garnered significant public attention, reflecting the government's strong emphasis on the humanoid robot industry. This focus is expected to accelerate the development of the humanoid robot sector and help overcome key challenges for commercialization [3] Popular ETFs - The Robotics ETF (562500) is the largest robotics-themed ETF in the market, facilitating investors' access to the Chinese robotics industry [3] - The Huaxia Sci-Tech AI ETF (589010) serves as the brain of robotics, with a 20% fluctuation range and small-cap elasticity, aiming to capture pivotal moments in the AI industry [3] Technological Developments - On April 18, Microsoft launched the BitNet b1.58 2B4T, the largest 1-bit AI model globally, featuring 2 billion parameters and capable of efficient operation on standard CPUs like Apple's M2. The model utilizes minimal weights of -1, 0, and 1 to achieve high memory and computational efficiency, outperforming similar models from Meta and Google in various inference tasks [2]

其中，夺冠的天工Ultra背后是港股上市公司优必选。2024年全年，优必选实现营收12.95亿元，净利润 则亏损11.24亿元。其他公司则均未上市。 新华财经北京4月21日电（可达）世界首次"人机共跑"半程马拉松赛事4月19日在京完赛，来自北京人形 机器人创新中心的"天工队"以2小时40分42秒的成绩摘得冠军，"松延动力小顽童队"和"行者二号队"分 获亚军、季军。 除了"看热闹"，本次赛事背后有哪些厂商入局？除了机器人本身，还和资本市场摩擦出了哪些火花？ 优必选成赢家还有哪些公司亮相？ 据悉，本次有国内高校、科研机构、企业等20支机器人队伍参赛。其中，有6支队伍成功完赛，分别为 天工队、小顽童队行者二号队、旋风小子队、EAI队、半醒机器人队。值得注意的是，只有天工队在没 有更换机器人的情况下完赛，其余队伍最多的中途更换了五次机器人。 | 机器人名称 | 劇海 | 应用 | 所属以在 | | --- | --- | --- | --- | | 天工Ultra | 100cm | 52kg | 天工队 | | TI70A | 1.7m | 1.7m | は虎鍋 | | 神农 | 173cm | 50kg | 神农 ...

央视网消息：全球首个人形机器人半程马拉松胜利收官，20个不同型号的人型机器人各具特色，引燃现场气氛。从跑赢赛事到跑 通产业链，这场人形机器人半程马拉松比赛其实也是一次对人形机器人行走、奔跑、全身协同、具身智能等能力的全方位检验。人形 机器人产业发展的"马拉松"才刚刚开跑。 北京经济技术开发区管理委员会副主任刘力介绍，人形机器人半马顺利完赛，对机器人产业有很好的带动和引领作用，迅速的形 成样机，从样机有很好的供应链体系的供应。推进机器人进工厂，机器人进学校，机器人进家庭，用这些需求来带动创新研发和制 造，形成完整的产业生态。 在半马赛场上，成绩并非唯一焦点。机器人在奔跑中更换电池、有的机器人需要给关节喷制冷剂降温、有的则用上了仿生面 孔……这些细节都吸引着外界关注。赛事背后，高密度电池、一体化关节、智能传感设备等关键技术正在加速成熟，展现了我国在人 形机器人这一新兴技术交叉融合的"竞争高地"上的综合实力。 从实验室迈向产业化 人形机器人应用场景逐步落地 业内普遍认为，今年将成为人形机器人量产元年，从实验室逐步迈向产业化。此前有数据预测，2026年中国人形机器人产业规模 有望突破200亿元。专家表示，人形机器人主 ...