True

10月12日到16日，第二十六届中国国际工业博览会在国家会展中心（上海）举行。思灵机器人Agile Robots这回带着12款展品回来，把具身智能、具身模型、科研生态、消费应用、工业落地五个展区一起铺开，系统呈现从人形机器人、模型、力控与感知技术、数据采集体系，到多元场景解决方案的全栈能力。

From October 12 to 16, the 26th China International Industry Fair ran at the National Exhibition and Convention Center in Shanghai. This time Agile Robots returned with 12 exhibits, spreading out five zones, embodied intelligence, embodied models, the research ecosystem, consumer applications, and industrial deployment, presenting a full-stack capability that runs from humanoid robots and models through force control and sensing, data acquisition, and multi-scenario solutions.

思灵的底气，在于对这条技术链的长期投入。从电机、减速器、传感器，到机械臂、灵巧手和人形机器人，再到机器人操作系统、基础模型与应用算法，它构建了一套覆盖核心软硬件的全栈能力，并把它一直延伸到产线和工厂级应用。全栈的意义，不是把更多技术塞进一台设备，而是能围绕一个具体任务协同优化，让模型理解的任务变成硬件能执行的动作，让接触反馈及时进控制闭环，让现场的问题反过来推着产品迭代。

Agile's confidence sits in its long investment across the chain. From motors, reducers and sensors, to arms, dexterous hands and humanoid robots, then to the operating system, foundation models and application algorithms, it has built a full-stack capability spanning core software and hardware, and extends it down to production lines and factory-level applications. The point of full-stack is not to cram more tech into one machine, but to co-optimize around a concrete task: turn the task a model understands into an action the hardware can execute, let contact feedback enter the control loop in time, and let on-site problems push the product to iterate.

思灵机器人创始人兼CEO陈兆芃对这件事的界定很直接，具身智能并不是简单地把大模型装进一个机器人身体。它需要模型、躯体、执行机构、感知系统和应用场景的协同，更需要通过高价值应用的大规模部署，形成一个应用、数据、模型、硬件、再应用的持续循环。

Chen Zhaofan, co-founder and CEO of Agile Robots, defines it plainly. Embodied intelligence is not simply stuffing a large model into a robot's body. It requires the coordination of model, body, actuators, sensing and application scenarios, and, above all, large-scale deployment of high-value applications to form a continuous cycle of application, data, model, hardware, and back to application.

2026年，首届EBiM，也就是具身双臂操作挑战赛正式启动，吸引了全球119支团队报名。赛事要建的，是一套面向真实世界的具身智能标准化评测体系，让机器人算法不仅能在仿真里有效，还能在不同实验室、不同环境、不同真实机器人平台上得到验证。作为主要赞助商，思灵为比赛提供了统一平台Mobile FR3 Duo，让全球研究者能拿算法上真实硬件去接受检验。

In 2026, the first EBiM, the Embodied Bimanual Manipulation Challenge, officially launched, drawing 119 teams worldwide. The event is building a standardized evaluation system for embodied intelligence aimed at the real world, so robot algorithms are not merely effective in simulation but verified across different labs, environments and real robot platforms. As a principal sponsor, Agile supplies the unified platform, the Mobile FR3 Duo, so global researchers can run their algorithms on real hardware to be tested.

展品里有一台摄影机器人，把AI、力控和拍摄工作流绑在一起，让摄影师用自己熟悉的创作逻辑去使唤它。跟工业机器人复杂的示教操作不一样，这台支持拖拽打点，用户直接把机械臂引导到目标位置、设定关键帧，剩下的运镜路径交给轨迹规划模型去生成。

Among the exhibits is a photography robot that binds AI, force control and the shooting workflow, so a photographer can command it with familiar creative logic. Unlike the complex teach-operation of industrial robots, this one supports drag-and-drop marking; the user simply guides the arm to a target position, sets key frames, and lets a trajectory-planning model generate the camera path.

把这12款展品放在一起看，思灵在工博会上呈现的，不是一堆孤零零的产品，而是一套完整的产业化能力，持续创新的研发体系、面向物理世界的AI与机器人技术、可扩展的制造能力，以及连接全球市场的交付网络。

Set the 12 exhibits together and what Agile shows at the fair is not a pile of isolated products but a complete set of industrialization capability: a continuously innovating R&D system, AI and robotics aimed at the physical world, scalable manufacturing, and a delivery network that reaches global markets.