本报（chinatimes.net.cn）记者赵奕 上海报道 根据上交所官网，鉴于节卡机器人股份有限公司（下称"节卡股份"）尚有相关事项需进一步核查，因此 取消了该公司原定于8月8日的科创板IPO上会审议，这也让节卡股份成为年内首家上会前取消审议的企 业。 节卡股份与越疆科技、遨博机器人一同被称为"协作机器人三杰"，2024年12月，越疆科技成功登陆港 股，拿下"港股协作机器人第一股"之称，这也让市场对节卡机器人的IPO进展更加关注。不过，目前距 离节卡股份IPO受理已经超过两年，而其虽身处机器人产业发展风口，却依旧面临盈利能力不稳定等问 题。 针对本次IPO相关问题，《华夏时报》记者致函节卡股份，但截至发稿并未收到回复。 营收增长但盈利稳定性不足 根据公开信息，节卡股份成立于2014年，主要从事协作机器人整机产品的研发、生产、销售，并从事包 括集成设备及自动化产线在内的机器人系统集成业务。 在商业化方面，节卡股份实现了在"中国天眼"（FAST）运行维护、汽车水冷板焊接、航天装备部件装 配、半导体封测等高端技术场景的应用突破，目前已广泛应用于汽车零部件、3C电子、半导体、精密 制造、新能源等智能制造领域，并覆 ...

Core Points - The Future Science Prize 2025 was announced on August 6, 2025, in Beijing and Hong Kong, revealing the winners of the "Life Science Prize," "Material Science Prize," and "Mathematics and Computer Science Prize" [1] - The Future Science Prize, established in 2016, aims to recognize outstanding scientific achievements in basic research within mainland China, Hong Kong, Macau, and Taiwan, regardless of nationality [4] Life Science Prize - Winners: Ji Qiang, Xu Xing, and Zhou Zhonghe [5] - They were awarded for their discovery of fossil evidence supporting the theory that birds originated from dinosaurs, transforming a long-standing hypothesis into a widely accepted scientific theory [6][9] Material Science Prize - Winners: Fang Zhong, Dai Xi, and Ding Hong [13] - Their contributions include the computational prediction and experimental realization of topological electronic materials, which are considered one of the most significant advancements in condensed matter physics [15] - Fang Zhong and Dai Xi developed a comprehensive set of computational methods to predict various topological materials, while Ding Hong played a crucial role in experimentally verifying the existence of Weyl fermions in predicted materials [15] Mathematics and Computer Science Prize - Winner: Lu Zhiyuan [23] - Recognized for his inventions and leadership in the fields of non-volatile semiconductor memory unit density, device integration, and reliability [25] - Lu Zhiyuan pioneered next-generation non-volatile memory (NVM) technologies, which have significant implications for artificial intelligence, mobile communications, cloud computing, and edge computing [26]

Group 1: Economic Development in Guangdong - The integration of digital intelligence is accelerating the upgrade of industries such as food, trendy toys, and petrochemicals in Guangdong, transitioning from traditional manufacturing to intelligent manufacturing [1][4] - Data from China Mobile Guangdong indicates that from January to June 2025, the average number of permanent staff in 96 key industrial transfer parks reached 1.273 million, an increase of 43,000 year-on-year, reflecting a growth rate of 3.5% [1] Group 2: Technological Innovations - AI, low-altitude, and robotics technologies are reshaping the structure of social productivity and are key drivers for industrial upgrading and high-quality development [2] - The implementation of "cloud network + AI + low-altitude" technology at Maoming Petrochemical has enabled significant operational improvements, achieving a 15-fold increase in inspection efficiency and a 92% reduction in single operation costs [3] Group 3: Industry-Specific Advancements - The introduction of 5G and intelligent forklifts at Donggu Food Company has significantly improved warehouse management efficiency, allowing for precise cargo location and one-click order completion [4] - The trendy toy industry in Dongguan is experiencing a surge in orders, necessitating enhanced collaboration between design, production, and sales, supported by high-speed dedicated networks [4] Group 4: Growth in New Industries - In the first half of the year, Guangdong's advanced manufacturing and high-tech manufacturing sectors saw industrial added value growth rates of 5.9% and 6.0% respectively [5] - The production of industrial robots, service robots, and civilian drones has seen substantial growth, with increases of 34.0%, 23.0%, and 58.2% respectively, positioning Guangdong at the forefront nationally [5][6]

近年来，人工智能技术持续向纵深发展，以智能体为代表的大模型应用正成为推动产业智能化变革的核心引擎之 一。智能体可以将模型技术能力与任务执行能力充分结合，高度契合日益复杂的提质增效需求，有望成为推进人 工智能与实体经济深度融合的关键力量。 2025年7月28日上午，世界人工智能大会（WAIC）智能体驱动产业变革论坛成功举办。会上，中国人工智能产业 发展联盟副秘书长、中国信息通信研究院（简称"中国信通院"）副总工程师王爱华正式发布并解读"2025智能体 十大关键词"，十大关键词分别是通用智能体、专用智能体、端侧智能体、企业级智能体、智能体平台、智能体通 信协议、多智能体协同、智能体价值评估、智能体可信、智能体基准测试，集中体现了当前智能体未来发展趋 势、技术研究热点与产业关注重点。 一、通用智能体：面面俱到的"数字通才" 通用智能体具备跨领域知识和通用能力，能够理解并执行多种类型任务，是迈向通用人工智能的重要路径。通用 智能体的发展不仅依赖大模型的能力提升，也需要感知规划、认知决策、多模态融合、人机协作等多技术融合创 新。未来，通用智能体将沿着多阶段能力跃迁与多维度技术融合的方向发展，驱动人机协同从"被动响应"迈 ...

Core Viewpoint - FAIR plus is a platform focused on the entire robotics industry chain, aiming to promote collaboration, technology exchange, and resource matching within the robotics and artificial intelligence sectors [6][10]. Group 1: Event Overview - The FAIR plus 2026 event is scheduled for April 22-24, 2026, at the Shenzhen Convention and Exhibition Center, with an exhibition area of 15,000 square meters [19]. - The event will feature a combination of academic conferences, technical salons, community cultivation, and supply-demand matching [19]. Group 2: Objectives and Vision - The event aims to facilitate deep dialogue between upstream and downstream sectors of the industry, discover business opportunities, and promote collaborative development [10]. - It seeks to advance the integration of artificial intelligence and robotics, enhancing China's robotics industry capabilities on a global scale [10]. Group 3: Historical Context and Achievements - The previous FAIR plus 2025 event set new records for the Shenzhen robotics exhibition scale, attracting over 200 companies from the robotics industry chain, with three-quarters being component suppliers [21]. - FAIR plus 2025 hosted 23 supply-demand matching sessions, facilitating the application of robotics in various real-world scenarios [22]. Group 4: Exhibition Features - The exhibition will cover over 50 segments of the embodied intelligent robotics industry, showcasing more than 500 upstream and complete machine enterprises [13]. - Special zones will include areas for startups, showcasing innovative products and business models, as well as a research institution innovation display area featuring original research outcomes [14][15]. Group 5: International Collaboration - The event will open an international exhibition area to facilitate deep exchanges and cooperation between domestic and foreign robotics technologies, standards, and applications [16]. - Over 100 overseas professional buyer groups are expected to attend, enhancing global engagement in the robotics sector [13].

Core Viewpoint - Tencent is entering the embodied intelligence trend by developing a modular external brain for robots, rather than engaging in hardware production or commercialization [2][8][22]. Group 1: Tairos Platform Overview - The Tairos platform serves as a general external brain for robots, allowing various manufacturers to access specific capabilities without needing to develop their own hardware [3][10]. - It integrates Tencent's software capabilities in embodied intelligence, including multi-modal perception, planning, and action models [6][12]. - The platform is designed to be modular, enabling robot manufacturers to select and utilize the components that best fit their needs [16][20]. Group 2: Model and Cloud Services - The core of the Tairos platform consists of three main models: multi-modal perception model, planning model, and perception-action joint model, each analogous to parts of the human brain [12][13]. - The platform provides cloud services in a tool platform format for developing, training, testing, and deploying robotic applications [14][17]. - It allows for flexible integration, where manufacturers can use specific models from Tairos to enhance their robots' capabilities [18][19]. Group 3: Strategic Focus and Industry Position - Tencent emphasizes its "three no principles": not producing hardware, not engaging in mass production, and not pursuing commercialization [22][23]. - The company aims to focus on foundational research and exploration in embodied intelligence, rather than limiting itself to hardware development [34][35]. - By collaborating with various robot manufacturers, Tencent seeks to gain insights into industry trends and challenges, enhancing its software offerings [37][39]. Group 4: Future Directions - The future development of embodied intelligence will revolve around the "IDEAS" framework, which includes virtual-real integration, lowering technical barriers, intelligent evolution, agentification, and perception expansion [39][40]. - Tencent's goal is to accelerate the industry towards a pivotal moment in embodied intelligence, akin to the introduction of the iPhone [40][41].

机器人产业蓬勃发展，正极大改变着人类生产和生活方式，为经济社会发展注入强劲动能。2023年年初，工业和 信息化部联合十六大部委印发《"机器人+"应用行动实施方案》，目标到2025年，制造业机器人密度较2020年实现 翻番，服务机器人、特种机器人行业应用深度和广度显著提升，机器人促进经济社会高质量发展的能力明显增 强。11月12日工业和信息化部又印发《人形机器人创新发展指导意见》，提出人形机器人有望成为继计算机、智 能手机、新能源汽车后的颠覆性产品，并明确2025年实现批量生产。 各项政策加持，热度非凡，行业活动也层出不穷，小编通过网络搜集整理了 未来两周国内外机器人及人工智能领 域行业活动活动榜单 ，供大家参考。 1 第40届中国计算机应用大会（CCF NCCA 2025） 时间： 2025.08.07-08.09 地点： 北京·中国科技会堂 主办单位： 中国计算机学会（CCF） 2025.08.04-08.17 机器人及人工智能相关活动预告 2 第四届先进制造技术与制造系统国际学术会议(ICAMTMS 2025) 时间： 2025.08.08-08.10 地点： 安徽芜湖市 3 第二届无人农场与智能装备应用 ...

以下文章来源于机器人技术与应用 ，作者唐 机器人技术与应用 . 传播企业信息和市场行情，交流业内创新成果，推动行业技术进步。宣传报道国内外机器人技术领域最新技术、 成果和信息，促进企业转型升级，搭建产学研交流平台。 机器人，正在 "听懂" 人类。 从需要写代码到 "说句话、动动手" 就能教，机器人训练方式的变化，其实是 "人机协作" 的一大步。 1. 远程操控模式 ： 通过操纵杆控制机器人完成高危任务（如处理有毒物质）； 长期以来，训练机器人掌握新技能是编程专家的专属领域，但 MIT工程师最新研发的 多功能演示界面（ VDI） 将 彻底改写这一规则。这款手持式智能工具首次整合 远程操控、运动引导、自然模仿 三大训练模式于一体，让普通 工人、护理人员甚至家庭成员都能成为机器人的 "老师"。 简单说，它是一个装了摄像头、力传感器和标记点的 "万能手柄"，插在任何常见协作机械臂上即可使用。 这种方式的创新点： | 传统做法 | 新做法 (VDI) | | --- | --- | | 写程序 → 调试 → 再写程序 | 直接上手演示 | | 一种训练方法用到黑 | 3 种方法随时切换 | | 只有工程师能操作 | 工 ...

Core Insights - The article emphasizes the deep integration of artificial intelligence and robotics, driving embodied intelligent agents towards higher levels of autonomy and intelligence [2][12] - It highlights the importance of core components in embodied intelligent robots, including high-precision sensors, bionic actuators, edge computing chips, and flexible materials, which directly impact overall performance and application scenarios [2][13] Forum Focus - The forum will focus on the technological frontiers of core components for embodied intelligence, common challenges such as reliability, cost control, and standardization, and application ecosystems in humanoid robots, industrial automation, and medical rehabilitation [2][13] - The event is organized by several institutions, including the China Association for Science and Technology and Beijing Institute of Technology, indicating a strong institutional backing [2][11] Event Details - The forum is scheduled for August 10, 2025, from 14:00 to 17:30 at the Beijing Yichuang International Exhibition Center [3][11] - A detailed agenda includes various presentations from industry experts on topics such as data-driven robotics, innovations in bionic actuators, and the role of synthetic data in embodied intelligence [4][15] Keynote Speakers - Notable speakers include Zhang Jianwei, a foreign academician of the Chinese Academy of Engineering, and Jiang Lei, chief scientist at the National Collaborative Innovation Center for Humanoid Robots [4][15] - The event will also feature discussions on the integration of physical models and data-driven approaches in robotics, showcasing the latest advancements in the field [4][16] Networking Opportunities - The forum will provide opportunities for networking and collaboration among industry leaders, researchers, and companies, aiming to build a complete ecological loop from innovation to industrial application [6][10]

Core Viewpoint - The core discussion revolves around the concept of embodied intelligence, emphasizing the intricate relationship between body, environment, and intelligence, and how these elements collectively contribute to the realization of intelligent systems [4]. Group 1: Embodied Intelligence - Embodied intelligence is defined by three key elements: body, environment, and intelligence, which interact in complex ways to enable intelligent behavior [4]. - The structure and sensory capabilities of the body significantly influence how an intelligent agent perceives and interacts with the world, highlighting the importance of physical attributes such as height and limb structure [4]. Group 2: Large Models in Embodied Intelligence - The training of embodied large models requires the integration of visual, linguistic, and behavioral data, necessitating a unified approach to data, computing power, and algorithms [4]. - The complexity of data in training embodied large models is heightened as it must encompass multimodal information, including behavior, physical parameters, and tactile data [4]. - Challenges remain in the generalization capabilities of embodied large models in real physical spaces, particularly concerning data complexity and sensor differences [4]. Group 3: World Models - World models serve as abstract representations of the real world, encompassing three-dimensional space, dynamic changes, object relationships, and memory, which are crucial for understanding and predicting environmental states [5]. - The relationship between world models and large models, as well as their connection to three-dimensional spaces, presents areas for further exploration [5]. - Current research often relies on simulators to generate data, but aligning virtual environments with real-world physical parameters remains a significant challenge [5].