2026年1月31日，以"向新向优，笃行致远"为主题的第四届江苏财经年会在南京举行。政、产、学、研各界 精英汇聚一堂，共同回望2025年江苏经济脉络，展望"十五五"发展新程。 会上，数字华夏创始人、首席执行官沈健获评"2025江苏财经新闻人物"。 这位从江苏走出的创业者，在投身人形机器人领域仅一年多后，便带领企业签下数亿元订单，其产品已进入 银行、学校等多个核心场景。与众不同的是，沈健从一开始就锚定了一条差异化的路径，自研突破机器人智 能交互的技术瓶颈，解决机器人与人之间沟通难题，打造"有温度的人形机器人"，领跑具身交互智能产业。 在硅胶皮肤与精密算法的背后，他主导的这场关于人机交互的产业实践渐行渐深。 让机器人完成"十小时表情测试" 01. 在南京和上海的研发中心，数字华夏的人形机器人"夏澜"正在升级，作为第三代仿生机器人，她需要进行一 项苛刻的测试：保持自然、和谐的表情连续工作十小时。关键在于确保每一个微笑、每一次眼神流转在长时 间运行下都不失真、不僵硬。 "让机器人展示'温度'，依赖算法、算力和数据，这必须建立在稳定与可靠之上。"沈健拆解这一测试的目 的，在高度拟人的硅胶面庞之下，包裹着的是一颗经过特殊 ...

Core Insights - The commercialization of humanoid robots is driving the demand for advanced motor technologies, which are essential for performance, load capacity, and operational efficiency [1][3][15] - The electric motor industry is becoming a critical variable in the speed of commercialization for humanoid robots, with significant competition expected by 2026 [1][3] Electric Motor Market Potential - The global market for frameless torque motors in humanoid robots is projected to reach $2.397 billion by 2030, with the total market exceeding $3.3 billion when including traditional sectors [3] - Hollow cup motors are expected to surpass $1.5 billion in the robot sector by 2030, while harmonic field motors are anticipated to grow rapidly post-2025, reaching $2 billion by 2030 [3] - Electric motors account for 15%-20% of the value chain in humanoid robots, prompting leading manufacturers to vertically integrate and customize their motor solutions [3][6] Performance Requirements - Humanoid robots require motors with high torque density, low torque fluctuation, strong overload capacity, and effective heat dissipation [4][6] - For instance, the A2 Max robot's joint peak torque reaches 450 Nm, necessitating high-density output in a compact space [4] - The design trends are moving towards modular and integrated solutions to enhance flexibility and endurance [4] Competitive Landscape - The current electric motor market features a multi-route approach with frameless torque motors leading for joints, hollow cup motors for dexterous hands, and harmonic field motors exploring future applications [6][7] - The global market is dominated by companies like Kollmorgen and AEROTECH, while domestic firms are rapidly advancing, with significant price advantages [7][10] Market Growth and Trends - The global compound annual growth rate (CAGR) for frameless torque motors is projected to be 25% from 2023 to 2030, with the market expected to exceed $3.3 billion by 2030 [7][10] - Domestic companies are making strides in technology and cost efficiency, with local products priced at 50%-70% of their international counterparts [10][11] Material Innovations - The fourth-generation samarium iron nitrogen (SmFeN) material is emerging as a game-changer in electric motor performance, offering low cost and high performance [11][12] - This material has a Curie temperature of 470°C and is resistant to corrosion, making it suitable for micro-motor applications [12][14] Future Outlook - The humanoid robot wave represents not just a competition among complete systems but also a "power revolution" in core components like electric motors [15][16] - As technology converges and the supply chain integrates, the electric motor industry is expected to become a key driver for the commercialization and popularization of humanoid robots [16]

2026 年，中国具身智能产业在政策与资本助推下活力凸显，落地场景持续拓展，头部企业已规划万台级产能。然而行业亦面临严峻挑战：多数企业仍难突破百台 量产瓶颈，受制于关键零部件供应不足、单台成本过高及成熟应用场景有限等现实问题，呈现出"冰火两重天"的发展态势。 面对规模化交付的核心命题与全方位考验，产业亟需一个汇聚智慧、共享经验、推动协同的关键平台。在此背景下， 第三届中国具身智能与人形机器人产业大会将 于 4 月在京启幕 ， 旨在集行业之力， 共同 解答 产业规 模化命题。 本届大会以 "竞逐具身人形万亿赛道·重塑未来产业新纪元" 为主题，将汇聚全球院士专家、企业领袖、投资人、产业链上下游企业及场景方，通过 【大会论坛 +展 览展示+颁奖典礼+报告发布+供需对接会+人才对接会】 多元模式，打造产业生态共建、资源共享、合作共赢的高能级平台，为培育新质生产力的核心增长极注入 强劲动能。 ▍擘画六大价值版图，驱动企业全周期成长 01 板块一 主论坛顶尖声音，定义产业未来 当前，人形机器人产业正处在一个激动人心却又充满不确定性的十字路口。在百家争鸣、尚无定 论的关键时期，聆听最前沿、最权威的思想声音，是拨开迷雾、锚定 ...

一个瓶盖在机械手指间被旋转、拧开，整个过程流畅得仿佛人手。但驱动这套动作的眼睛只是一个 普通 摄像头，皮肤是成本不到 250 美元的 触觉 传感器。 长久以来，机器人专家深信不疑：要复现人手无与伦比的灵巧性，必须配备与之匹配的、同样精密的感知系统 ， 如 高分辨率视觉、能绘制精细力场分布图的触觉 传感器 。 然而， 来自浙江大学等机构研究团队发表在 《 Science Robotics 》的最新研究 《 Visual-tactile pretraining and online multitask learning for humanlike manipulation dexterity》 ， 以其彻底的降维方式给出颠覆性的答案。 研究团队证明，仅依靠单目 RGB 摄像头和提供二值触觉信号的简易 触觉 传感器， 机器人手就能在拧瓶盖、转水龙头、滑动杠杆等复杂任务中达到约 85% 的成 功率，并能举一反三，完成削铅笔等未训练任务。 更令人惊讶的是，这套感官系统的总成本仅约 250 美元。 怎么实现的？一套逻辑 ：灵巧的关键或许不在于感官的高保真，而在于大脑如何理解 并 融合 " 感 知 " 与 "控制" 。 ▍ ...

什么样的机器人能真正走进千家万户？《超能陆战队》里的大白机器人展示了一种可能—— 柔软、安全、会 拥抱 。 现在， 南 方科技大学王宏强教授团队 带来了现实版本 ： 一个受人体骨骼启发的 可生长、可变形、可灵活移 动的软体人形机器人 。 这个名为 GrowHR 的机器人，拥有一系列特殊技能。 比如，它可能是第一个不怕被扔进泳池的人形机器人 ——不仅不会 "淹死" ，还能悠哉悠哉地游泳： 还能施展 "轻 功水上漂"， 在水面上稳稳行走： 厉害的是，这家伙能上 天，远程空投作业不是梦 ： 遇到狭窄空间？ "缩骨功"一开，轻松穿过： 弹性腿部还能蓄力踢球： 最关键的是，它足够安全，一个 6岁小孩都能轻松举起它 ，即 使发生跌倒碰撞 也不怕： 它的才艺可不止这些。 近日，相关研究 登上 国际 顶级期刊 Science A dva nces， 为更安全、更多功能的 人形机器人走入工厂或家庭等场景提供了全新的解决思路。 ▍ 仿生骨骼：能屈能伸的秘密 传统人形机器人为什么难以普及？ 质量太大、结构太刚硬是主要原因 。它们容易撞到人或周围环境，造成人 身和财产风险，机器人自身也容易被撞坏。 研究 团队从人体骨骼找到了灵感。人 ...

2026 年 1 月 29 日，劳尼克斯在上海举办生态合作伙伴大会。会议围绕智能园林工具的发展趋势、电动化进 程以及渠道协同模式展开交流，来自全国多地的经销商及生态伙伴参与其中。大会聚焦高性能锂电在园林工具 领域的应用方向，以及面向区域市场的长期合作与共建逻辑。 ▍ 园林工具迈入结构性转型阶段 园林工具行业正处在由传统燃油产品向电动化、低碳化转型的关键阶段。长期以来，燃油产品在作业能力与稳 定性上的优势，使其在市场中占据主导地位，但随之而来的高噪音、高排放及高维护成本问题，也逐渐成为行 业发展的制约因素。 随着电动化技术的持续成熟，行业关注点正在从 "是否可用"转向"是否真正适用于专业场景"。在这一过程中， 仅依靠低成本竞争的锂电产品难以建立长期优势，高性能锂电与智能化能力，正在成为推动行业升级的重要方 向。同时，渠道端也在重新审视自身的发展模式，更加关注经营的稳定性与可持续性。 ▍ 以电动化为基础的长期发展路径 在行业结构性变化的背景下，劳尼克斯将园林工具业务定位为以电动化为基础、以智能化能力为延展的长期发 展方向。公司方面认为，具备长期市场价值的电动化产品，应在动力表现、使用体验与系统稳定性之间实现平 衡 ...

2026 年 1 月 23 日，印度浦那的 Muks Robotics 正式宣布，自主研发的社交人形机器人 SPACEO-M1 开放全球企业订购，这款产品也成为印度首款实现量产 并推向全球的社交人形机器人。 该机器人早在 2025 年 3 月已启动预售，彼时仅面向印度国内客户，并非正式的全球开售；时隔9个月，量产版 SPACEO-M1 人形机器人终于迎来全球市场的正 式亮相。 ▍ Muks Robotics 创始人为印度牙科医生转型 Muks Robotics 由 Mukesh Bangar 博士于 2021 年创立。与行业内科班出身的创业团队不同，Bangar 博士早年毕业于马哈拉施特拉邦健康科学大学，持有牙 科学士学位，曾任职于 Chowgule Mediconsult Pvt. Ltd. 担任牙医。该公司初创阶段的核心团队成员均为跨领域技术爱好者，无传统机器人头部企业从业背 景，纯粹因对机器人领域的热爱凝聚成为初始创业团队。 Bangar 博士本职为牙医，长期关注医疗自动化设备领域，因洞察到公共服务场景中智能交互设备的供给缺口，他自主研习人工智能与机器人技术，跨界投身机 器人研发。其早期并未规划涉足人形 ...

Group 1 - The article discusses the development of a flying umbrella by Canadian engineer John Tse, which autonomously hovers above the user to provide rain protection [1][5] - The umbrella utilizes drone technology, featuring four propellers for lift and a depth camera for tracking the user's head position, allowing it to maintain a stable hover [7][9] - Despite its innovative design, the umbrella's practicality is questioned, as it must hover several meters above the user, resulting in limited rain coverage [5][20] Group 2 - The second-generation flying umbrella improves upon the first by eliminating the need for manual remote control, which was criticized for being cumbersome [7] - The depth camera used in the umbrella can function effectively in low-light conditions, enhancing its tracking capabilities compared to standard cameras [9][12] - The umbrella's design includes a foldable mechanical arm structure, allowing it to be compact and portable when not in use [11][14] Group 3 - The development process involved numerous challenges, including hardware failures and software issues, extending the project timeline to nearly a year [18] - The umbrella's current battery life is limited to 10-15 minutes, similar to small consumer drones, raising concerns about its usability in crowded areas [20] - The project highlights the potential of personal creators to leverage existing drone technology and open-source hardware to create innovative solutions, even if they are not immediately practical [20]

沃庭科技是一家全球化布局的技术创新型科技公司，专注于通过人工智能与先进硬件技术重塑电动工具体验。 公司以自主研发的 20V 智慧锂电工具生态为核心，致力于打造覆盖全场景、智能化的一站式工具解决方案， 为全球用户提供更安全、更高效、更具智能洞察的专业工具体验。目标成为电动工具行业最快崛起的新生力 量，打造全球化的电动工具高端品牌。 ▍ 围绕产品生态系统化建设，持续拓展产业级合作网络 上海， 1 月 29 日 —— 沃庭科技在上海举办 2026 电动工具生态合作伙伴大会。围绕电动工具产业的锂电 化、智能化与系统化演进趋势，沃庭科技在大会上集中展示其以 20V 智慧锂电电池包平台为基石的智慧锂电 工具产品生态体系，并同步释放在技术研发、产业协同及区域资源等多个维度的战略合作进展。 当前，电动工具行业正从单品性能竞争，逐步转向生态构建能力、系统稳定性及智能化协调能力的综合比拼。 沃庭科技此次大会，重点围绕这一命题，对外系统性呈现其产品与技术路径。 ▍ 以构建生态为核心，推出全场景智慧锂电工具 大会现场，沃庭科技集中展示了其围绕 20V 智慧锂电电池包平台所构建的产品生态体系。首批产品覆盖钻、 扳、锤、磨、锯等六大类核 ...

Core Viewpoint - Ant Group's Lingbo Technology has launched a comprehensive open-source initiative, unveiling four advanced models that enhance embodied intelligence, covering perception, decision-making, simulation, and execution [1][21]. Group 1: Vision Model - LingBot-Depth - LingBot-Depth addresses the challenge of transparency in robotic vision, enabling robots to "see the invisible" by overcoming limitations of traditional depth cameras [1]. - The model employs a Masked Depth Modeling (MDM) algorithm, enhancing existing sensors' capabilities without requiring expensive hardware upgrades, achieving a 47% reduction in error for sparse depth completion tasks and over 70% in indoor scenarios [2]. - In practical tests, robots using LingBot-Depth improved their success rate in grasping transparent objects from 0% to 50% [3]. Group 2: Brain Model - LingBot-VLA - LingBot-VLA utilizes 20,000 hours of real robot operation data, significantly enhancing the model's common sense and generalization capabilities, setting a new record in the open-source domain [5]. - The model's architecture incorporates a mixed Transformer design, integrating depth information from LingBot-Depth, resulting in an average success rate of 17.3% in real robot evaluations, surpassing the previous best [7]. - In the RoboTwin 2.0 simulation benchmark, LingBot-VLA outperformed the leading model by nearly 10 percentage points under highly randomized conditions [10]. Group 3: Simulation Model - LingBot-World - LingBot-World creates a high-fidelity, interactive simulation environment that adheres to physical laws, allowing robots to conduct zero-cost simulations and learn skills [13][17]. - The model can generate long videos without stability issues and allows real-time interaction, enabling users to manipulate the environment through natural language commands [14]. - This environment facilitates extensive trial and error, providing a platform for robots to learn and adapt before applying skills in the real world [17]. Group 4: Execution Model - LingBot-VA - LingBot-VA is the first self-regressive video-action integrated world model, enabling robots to "act while predicting" by generating the next frame and corresponding actions simultaneously [18]. - It features long-term memory capabilities, allowing it to distinguish between similar states in tasks, thus avoiding repetitive errors [20]. - The model demonstrates rapid learning with minimal demonstrations, achieving a 20% improvement in success rates for complex tasks compared to top baseline models, and reaching over 90% success in dual-arm collaborative operations [20]. Conclusion and Future Outlook - Ant Group's open-source initiative addresses the fragmentation and non-standardization in the robotics industry, offering a "base + adaptation" paradigm that allows companies to quickly adapt models to specific scenarios with less data and lower computational costs [21]. - This strategic positioning emphasizes the development of scalable, reusable core capabilities rather than manufacturing physical robots, marking a significant step towards a collaborative future in embodied intelligence [21].