第26届中国国际工业博览会期间，西门子共发布20余款新品，其中13款为本土创新产品，同时宣布启用中国五家工厂，开放六类场景与生态伙伴共创具身智能。20余款新品里，13款兼顾性能与质价比，属于本土创新产品，发布仪式上正式亮相的新一批为7款，包括升级版PLC逻辑控制器、变频器和全新的伺服系统等。
During the 26th China International Industry Fair, Siemens released over 20 new products, 13 of which are locally innovated, and announced the opening of five Chinese factories with six scenario types for co-creating embodied intelligence with ecosystem partners. Among the over 20 new products, 13 balance performance and cost-effectiveness as locally innovated products; the newly unveiled batch at the launch ceremony included 7 items, including upgraded PLC logic controllers, frequency converters, and entirely new servo systems.
西门子大中华区总裁肖松表示，中国不仅是全球最重要的工业市场之一，更是未来工业AI的关键应用场。丰富的场景、快速演进的客户需求和完整的产业生态，为工业AI从概念验证走向规模化落地提供了独特土壤。面对这一机遇，西门子将持续释放全球技术、中国创新与中国速度形成的合力和优势。此次工博会发布的多款本土创新产品、首次开放的具身智能工厂应用场景，正是这一方向的集中体现。西门子方面提供的背景资料里，两家装备企业用上本土创新产品后，一家单台设备年节能约52万元，另一家节省近30%厂房面积、拿到超50条生产线订单。
Siemens Greater China President Xiao Song said China is not only one of the world's most important industrial markets but also a key application field for future industrial AI. Rich scenarios, rapidly evolving customer needs, and a complete industrial ecosystem provide unique soil for industrial AI moving from concept validation to scaled deployment. Facing this opportunity, Siemens will continue to unleash the combined strength and advantages formed by global technology, Chinese innovation, and Chinese speed. The multiple locally innovated products released at this Industry Fair and the first-time opening of embodied intelligence factory application scenarios are concentrated reflections of this direction. In background materials provided by Siemens, after two equipment enterprises adopted locally innovated products, one saved about 520,000 yuan in annual energy per unit, and the other saved nearly 30% of factory space and secured over 50 production line orders.
本土创新的盘子做了两年。2024年10月，西门子成立中国创新产品事业部，将产品定义权、决策权放在本地。截至今年9月，已累计发布40余款本土创新产品。截至2026财年第三季度，中国本土创新产品营收增长近30%，中国区数字化工业自动化订单同比增长17%。部门随之更名为SMART产品事业部，对应可规模化、本土适配、AI就绪、可靠、及时五要素。新品里也有AI。Eigen工程智能体今年7月在中国上市，此次更新新增多轴驱动配置能力，并整合汽车行业常用的SICAR标准，近三个月已服务数百家企业。智能制造软件负责人韩三丰称，Eigen可缩短2至5倍的工程设置时间，企业方面提供的资料显示，程序开发时长最多可缩短50%。
The locally innovated portfolio has been built for two years. In October 2024, Siemens established the China Innovation Products business unit, placing product definition and decision-making authority locally. As of September this year, it has cumulatively released over 40 locally innovated products. As of Q3 fiscal year 2026, China locally innovated product revenue grew nearly 30%, and China digital industrial automation orders grew 17% year-over-year. The department was renamed the SMART Products business unit, corresponding to five elements—scalable, locally adapted, AI-ready, reliable, and timely. New products also include AI. The Eigen engineering agent launched in China in July this year; this update adds multi-axis drive configuration capability and integrates the SICAR standard commonly used in the automotive industry, having served hundreds of enterprises in the past three months. Smart manufacturing software head Han Sanfeng said Eigen can "shorten engineering setup time by 2 to 5 times"; company materials show program development time can be reduced by up to 50%.
随着物理AI从概念验证走向产业应用，真实工业场景正成为具身智能能力训练、验证与迭代的关键基础。制造现场工况复杂、任务多变、节拍严格、安全要求高，来自真实产线的测试反馈与数据积累，是具身智能形成可复制、可规模化工业应用能力的重要前提。工博会现场，西门子宣布开放6类支撑柔性自动化、具备行业通用性的具身智能工厂应用场景，包括无序散料智能分拣与精准上料、物料容器自主搬运与流转、物料容器智能拆垛与码垛、工艺操作与质量检测、高精度螺丝拧紧、制成品灵活包装。通过在中国5家工厂开放上述实景生产环境，西门子期待联合生态伙伴开展多维度测试验证，挖掘具身智能应用的跨产线、跨场景迁移潜力，推动具身智能加快落地生产现场。
As physical AI moves from concept validation to industrial application, real industrial scenarios are becoming a key foundation for embodied intelligence capability training, validation, and iteration. Manufacturing sites have complex conditions, varied tasks, strict cycle times, and high safety requirements. Test feedback and data accumulation from real production lines are important prerequisites for embodied intelligence to form replicable, scalable industrial application capability. At the Industry Fair, Siemens announced the opening of six types of embodied intelligence factory application scenarios supporting flexible automation with industry generality, including disordered bulk material intelligent sorting and precise loading, material container autonomous handling and circulation, material container intelligent depalletizing and palletizing, process operations and quality inspection, high-precision screw tightening, and finished product flexible packaging. By opening these real production environments at five Chinese factories, Siemens expects to conduct multi-dimensional testing and validation with ecosystem partners, exploring embodied intelligence applications' cross-line and cross-scenario transfer potential, accelerating embodied intelligence deployment on production sites.
西门子这次动作的核心，是把真实工厂开放出来给具身智能公司做训练场。前面讨论过国网天津电力把两处电力作业现场改成机器人学校，也讨论过上海推动8000台人形机器人进工厂，还讨论过北京人形数据与训练基地。这些都是同一个趋势的不同表现，真实场景正在成为具身智能落地的基础设施。西门子的做法和这些略有不同，它开放的是自己在中国五家工厂的具体产线场景，包括分拣、搬运、拆垛、质检、拧螺丝、包装这六类高频、标准化程度较高的任务。这些任务正好对应前面讨论的灵巧手抓取、移动操作、柔性装配等具身智能的核心能力。西门子把这些场景开放出来，等于给具身智能公司提供了一个真实的、带节拍和安全约束的测试环境，比实验室和样板间更接近量产条件。
The core of Siemens' move is opening real factories as training grounds for embodied intelligence companies. Earlier we discussed State Grid Tianjin Electric Power converting two power operation sites into a robot school, Shanghai pushing 8,000 humanoid robots into factories, and Beijing Humanoid Data and Training Base. These are different manifestations of the same trend—real scenarios are becoming infrastructure for embodied intelligence deployment. Siemens' approach differs slightly—it opens specific production line scenarios at its five Chinese factories, including six types of high-frequency, relatively standardized tasks: sorting, handling, depalletizing, quality inspection, screw tightening, and packaging. These tasks precisely correspond to embodied intelligence core capabilities discussed earlier—dexterous hand grasping, mobile manipulation, and flexible assembly. By opening these scenarios, Siemens provides embodied intelligence companies with a real testing environment with cycle time and safety constraints, closer to mass production conditions than laboratories and model rooms.
