近日，重工业具身智能机器人企业帝尔博格（DareRobot）完成亿元级Pre-A+轮融资，本轮投资方包括复星锐正、清石资管、正熙资本、西高投、博众精工、苏创投等机构。融资资金将主要用于重工通用具身核心技术研发、机器人产品迭代及市场拓展。帝尔博格成立于2022年，主要面向军工、钢铁、海装船舶、能源装备等重工业领域，提供智能切割、打磨、焊接、涂装等机器人产品，重点解决复杂工件、多品种小批量生产等场景中的自动化难题。
Hardware Kr learned that heavy industry embodied intelligence robot company DareRobot recently completed a hundred-million-yuan Pre-A+ funding round, with investors including Fosun RZ Capital, Qingshi Asset Management, Zhengxi Capital, Xigao Investment, Bozhon Precision, and Suchuang Investment. Funds will mainly be used for heavy industry general embodied core technology R&D, robot product iteration, and market expansion. DareRobot was founded in 2022, mainly serving heavy industry fields including military, steel, marine shipping, and energy equipment, providing intelligent cutting, grinding, welding, and coating robot products, focusing on solving automation challenges in complex workpieces and multi-variety small-batch production scenarios.
与汽车、3C等标准化程度较高的制造业相比，重工业的工件规格、几何形态和加工工艺差异较大，传统工业机器人往往需要针对不同工件进行编程和调试，多数工序长期依赖人工操作。团队方面，帝尔博格核心成员来自清华大学、西安交通大学、西安电子科技大学、西北工业大学等高校，多数成员曾任职于业内头部机器人公司。创始人张本毕业于清华大学核研院，获硕士学位，曾在核工业、船舶、建筑机械、工程机械等领域从事工业控制和机器人技术开发及管理，主导或参与多个国家重大专项课题，并曾从零组建机器人研发团队。
Compared with manufacturing industries with higher standardization like automotive and 3C, heavy industry has greater differences in workpiece specifications, geometry, and processing techniques. Traditional industrial robots often require programming and debugging for different workpieces, and most processes have long relied on manual operation. On the team side, DareRobot's core members come from Tsinghua University, Xi'an Jiaotong University, Xidian University, Northwestern Polytechnical University, and others, with most having worked at leading robotics companies. Founder Zhang Ben graduated from Tsinghua University's Institute of Nuclear and New Energy Technology with a master's degree, and has worked in industrial control and robot technology development and management in nuclear industry, shipbuilding, construction machinery, and engineering machinery fields, leading or participating in multiple national major special projects, and having built a robot R&D team from scratch.
在一家大型钢铁企业的生产车间，金属棒材加工后需要对端面不规则的残留部分进行打磨。由于待处理区域的位置和形态并不固定，这项工作过去主要依靠工人手持工具完成。对于传统工业机器人或专机而言，此类任务的难点在于工件去除特征存在差异，难以直接采用固定的加工轨迹。企业还需要兼顾连续生产条件下的加工效率、质量一致性和设备稳定性。帝尔博格CEO张本告诉硬氪，客户最初对机器人能否完成这项任务持怀疑态度。经过样件测试和前期验证，双方开始推进机器人在实际产线中的应用。目前，帝尔博格智能打磨机器人在该客户的生产现场已稳定运行1年多，客户还有进一步采购计划。
In a large steel enterprise's production workshop, metal bars after processing need grinding of irregular residual parts on the end face. Because the position and shape of the area to be treated are not fixed, this work was previously mainly done by workers holding tools. For traditional industrial robots or special machines, the difficulty of such tasks lies in differences in workpiece removal features, making it hard to directly use fixed processing trajectories. Enterprises also need to balance processing efficiency, quality consistency, and equipment stability under continuous production conditions. DareRobot CEO Zhang Ben told Hardware Kr that the customer initially doubted whether robots could complete this task. After sample testing and preliminary validation, both parties began advancing robot application in actual production lines. Currently, DareRobot's intelligent grinding robot has been running stably at the customer's production site for over a year, and the customer has further procurement plans.
据张本介绍，机器人能够识别工件的待加工部位，自主规划打磨路径，并按照客户要求完成加工后的质量检测。在连续生产条件下，单台设备的作业效率约为人工的4倍。设备的经济性也是客户采购的重要考虑。按照张本提供的设备效率和人工成本测算，该项目的设备投资回收期可以缩短至数月。类似需求也存在于船舶、能源装备及其他大型装备制造领域。目前，帝尔博格已经形成覆盖坡口切割、打磨、焊接、涂装等工序的智能机器人产品体系，并在多个工业场景中开展应用。
According to Zhang Ben, the robot can identify the workpiece area to be processed, autonomously plan grinding paths, and complete post-processing quality inspection per customer requirements. Under continuous production conditions, a single device's operating efficiency is about 4 times that of manual labor. Equipment economics is also an important consideration for customer procurement. Based on equipment efficiency and labor cost calculations provided by Zhang Ben, the project's equipment investment payback period can be shortened to several months. Similar needs also exist in shipbuilding, energy equipment, and other large equipment manufacturing fields. Currently, DareRobot has formed an intelligent robot product system covering groove cutting, grinding, welding, coating, and other processes, applied in multiple industrial scenarios.
复杂制造场景对机器人的要求，并不局限于识别工件和完成指定动作。在传统工业自动化方案中，工程师通常需要根据工件的位置、尺寸和加工要求，预先编写或调整机器人程序。当工件发生变化时，设备可能需要重新示教或调试。在多品种、小批量生产中，频繁调整程序不仅增加工程实施人力成本，也会影响设备的利用效率。帝尔博格通过物理感知、数据学习、模型推理、自主路径规划和工艺控制，减少机器人对人工示教的依赖。以打磨场景为例，机器人首先通过三维视觉获取工件及特征的空间几何信息，再通过AI算法模型输出目标区域运动轨迹与控制策略，综合几何轨迹、运动速度、力度、角度姿态、刀具换型等多重变量，最终驱动执行机构完成柔性生产操作。
Complex manufacturing scenarios' requirements for robots are not limited to identifying workpieces and completing specified actions. In traditional industrial automation solutions, engineers typically need to pre-write or adjust robot programs based on workpiece position, size, and processing requirements. When workpieces change, equipment may need re-teaching or debugging. In multi-variety small-batch production, frequent program adjustments not only increase engineering implementation labor costs but also affect equipment utilization efficiency. DareRobot reduces robot dependence on manual teaching through physical perception, data learning, model reasoning, autonomous path planning, and process control. Taking grinding as an example, the robot first acquires spatial geometric information of workpieces and features through 3D vision, then outputs target area motion trajectories and control strategies through AI algorithm models, integrating multiple variables including geometric trajectory, motion speed, force, angle posture, and tool changeover, ultimately driving actuators to complete flexible production operations.
帝尔博格目前主要通过真实重工场景积累AI模型数据。据介绍，公司数据来源包括客户（非涉密）生产现场的设备调试生产、自有工厂工件加工、以及样件打样测试。通过数据积累和模型训练，提升智能机器人对新任务的适应能力。对于重工业智能机器人企业而言，完成单个项目的技术验证，只是商业化过程中的一个阶段。不同客户的制造标准、工件规格、生产节拍和现场条件存在差异。如果每进入一个新场景，都需要定制开发和完善系统，企业的交付效率和规模化能力将受到限制。因此，帝尔博格产品技术的发展重点一直落在软硬件的可复用、模块化设计上。
DareRobot currently mainly accumulates AI model data through real heavy industry scenarios. According to the company, data sources include customer (non-classified) production site equipment debugging production, self-owned factory workpiece processing, and sample proofing tests. Through data accumulation and model training, it improves intelligent robots' adaptability to new tasks. For heavy industry intelligent robot enterprises, completing technical validation for a single project is only one stage in the commercialization process. Different customers have differences in manufacturing standards, workpiece specifications, production cycle times, and site conditions. If entering each new scenario requires custom development and system refinement, the enterprise's delivery efficiency and scaling capability will be limited. Therefore, DareRobot's product technology development focus has consistently been on reusable, modular software and hardware design.
以打磨机器人为例，机加件、下料件、铸件、锻件等行业工件特征存在明显差异，但其加工任务在感知识别、数据处理、轨迹规划及部分控制技术上具有共性。帝尔博格希望通过持续积累不同场景的工艺数据和算法能力，提高同类机器人对不同工件及生产要求的适应程度，以此实现跨行业客户的快速适配及交付。与此同时，公司正在推进DR_AI通用算法平台的建设。张本介绍，这一平台主要面向不同工艺机器人之间可以复用的基础技术，包括感知识别、数据处理和轨迹规划等能力。
Taking grinding robots as an example, machined parts, cut parts, castings, and forgings have differences in workpiece characteristics, but their processing tasks share commonalities in perception recognition, data processing, trajectory planning, and some control technologies. DareRobot hopes to continuously accumulate process data and algorithm capabilities from different scenarios to improve similar robots' adaptability to different workpieces and production requirements, achieving rapid adaptation and delivery for cross-industry customers. Meanwhile, the company is advancing construction of the DR_AI general algorithm platform. Zhang Ben introduced that this platform mainly targets reusable foundational technologies across different process robots, including perception recognition, data processing, and trajectory planning capabilities.
在产业资源布局方面，公司正在加强长三角地区、西南区域的业务布局，并同步建设海外市场渠道，以满足国内外重工客户的自动化需求。下一阶段，公司将进一步提高机器人处理大型、复杂结构工件的算法能力及技术通用性，拓展更多重工业制造场景。复星锐正表示，重工领域的智能化改造存在大量未被满足的真实需求，帝尔博格打通算法、机器人本体与重工工艺，已经跑出可复制的商业化落地路径。清石资管表示，帝尔博格在该领域有多年的机器人开发与交付经验，工艺覆盖完整，已与行业多家头部客户完成了量产交付。西高投表示，当下具身智能需要的不仅是宏大的叙事，更需要清晰的落地场景。
On industrial resource, the company is strengthening business布局 in the Yangtze River Delta and Southwest regions, while building overseas market channels to meet domestic and international heavy industry customers' automation needs. Next stage, the company will further improve robot algorithm capability and technical versatility for handling large, complex structural workpieces, expanding into more heavy industry manufacturing scenarios. Fosun RZ Capital said heavy industry intelligent transformation has unmet real needs; DareRobot connects algorithms, robot bodies, and heavy industry processes, having run out a replicable commercialization path. Qingshi Asset Management said DareRobot has years of robot development and delivery experience in this field, with complete process coverage, having completed mass production delivery with multiple leading industry customers. Xigao Investment said embodied intelligence today needs not just grand narratives but clear deployment scenarios.
