人形机器人的热度正在把产业链推到一个尴尬的位置。一边是具身智能公司不断刷新模型能力演示，另一边，真正决定机器人能否走出实验室的关节、传动与力控硬件，仍然卡在成本、寿命与柔顺性难以兼得的三角里。传统谐波减速器方案成熟，但价格高、结构占空间；直驱方案响应快，负载能力又常常跟不上。对于一批希望在具身智能早期阶段卡住底层硬件身位的公司来说，问题不是要不要做，而是能不能在量产之前，先把性能和成本账算到让本体厂商愿意切换的程度。
The heat of humanoid robots is pushing the supply chain to an awkward position. On one side, embodied intelligence companies continuously refresh model capability demonstrations; on the other, the joints, transmission, and force control hardware that truly determine whether robots can leave the laboratory remain stuck in a triangle of cost, lifespan, and compliance that is hard to satisfy simultaneously. Traditional harmonic reducer solutions are mature but expensive and space-consuming; direct-drive solutions respond fast but often cannot keep up in load capacity. For a batch of companies hoping to secure底层 hardware positions in the early embodied intelligence stage, the question is not whether to do it but whether, before mass production, they can calculate performance and cost accounts to the point where body manufacturers are willing to switch.
2026年10月20日，深圳具身智能操作平台与动力系统公司伊塔动力宣布完成超亿元种子轮和种子+轮融资。据机器人前瞻报道，本轮由经纬创投、商汤国香资本、建银国际、L2F光源创业者基金、银河通用、傅利叶智能联合投资。其中，银河通用与傅利叶智能以产业资本身份进入，使这笔交易在财务融资之外，多了一层产业链协同的意味。伊塔动力成立于2026年5月，距此轮融资披露仅约五个月。公司称其定位是具身智能硬件底层基础设施供应商，研发柔顺力控双臂与高性能关节模组。
On October 20, 2026, Shenzhen embodied intelligence operation platform and power system company Yita Dynamics announced completion of over 100 million yuan in seed and seed+ rounds. According to Robot Frontier reports, this round was jointly invested by Matrix Partners, SenseTime Guoxiang Capital, CCB International, L2F Light Source Entrepreneur Fund, Galbot, and Fourier Intelligence. Among them, Galbot and Fourier Intelligence entered as industrial capital, adding a layer of supply chain collaboration meaning beyond financial financing. Yita Dynamics was founded in May 2026, only about five months before this funding disclosure. The company says it positions itself as an embodied intelligence hardware底层 infrastructure supplier, developing compliant force-controlled dual arms and high-performance joint modules.
伊塔动力目前公开的产品矩阵是五款一体化关节模组，型号分别为YITA-50P-C、YITA-65P-C、YITA-84P-C、YITA-100P-C和YITA-122P-C。据公司披露，产品重量覆盖0.3kg至1.6kg，额定转矩覆盖2.2N·m至62N·m。公司称，其关节模组传动系统效率约为83%，成本较所对比的传统谐波方案降低约50%，硬齿面传动使用寿命超2万小时，约为传统方案的2至3倍。这是伊塔动力对外讲述的核心效率故事，也是其与谐波减速器方案正面竞争的主要依据。但需要明确的是，上述数据均来自公司单方声明，本次公开材料中未见第三方检测机构或客户侧的独立验证。
Yita Dynamics' currently public product matrix is five integrated joint modules—YITA-50P-C, YITA-65P-C, YITA-84P-C, YITA-100P-C, and YITA-122P-C. According to company disclosures, products cover 0.3 kg to 1.6 kg in weight and 2.2 N·m to 62 N·m in rated torque. The company says its joint module transmission system efficiency is about 83%, cost is about 50% lower than the compared traditional harmonic solution, and hard-tooth-surface transmission service life exceeds 20,000 hours, about 2 to 3 times traditional solutions. This is Yita Dynamics' core efficiency story told externally and its main basis for direct competition with harmonic reducer solutions. But it must be clear: the above data all come from unilateral company statements; no independent third-party testing or customer-side validation appears in this public material.
除关节模组外，伊塔动力还提供面向具身模型验证、机器人数据采集、末端执行器集成等场景的力控机械臂操作平台。在控制系统层面，公司称其围绕驱动器、通信协议和整臂控制器进行创新，结合负载补偿、力矩标定、阻抗控制及力位混合控制等技术，用于曲面打磨、柔顺抓取、双臂协同装配等操作任务。这意味着伊塔动力并不满足于只卖关节模组，而是希望以模组为入口，向上提供可被具身智能算法直接调用的操作硬件平台。
Beyond joint modules, Yita Dynamics also provides force-controlled robotic arm operation platforms for scenarios including embodied model validation, robot data collection, and end effector integration. On the control system level, the company says it innovates around drivers, communication protocols, and whole-arm controllers, combining load compensation, torque calibration, impedance control, and hybrid force-position control for tasks including curved surface grinding, compliant grasping, and dual-arm collaborative assembly. This means Yita Dynamics is not satisfied with selling only joint modules but hopes to use modules as an entry point to provide operation hardware platforms directly callable by embodied intelligence algorithms.
这里存在一个尚未被回答的关键问题，公司称成本较所对比的传统谐波方案降低约50%，但没有披露对比的具体型号、采购量级、精度等级或测试条件。关节模组的成本对比高度依赖对比基准，不同品牌、不同规格、不同批量的谐波模组价格差异显著。如果对比对象是进口高端谐波模组，50%的成本降幅与国产成熟谐波方案之间的实际价差可能完全不同。同样，83%的传动效率在行星方案中属于合理区间，但效率与精度、背隙、寿命之间存在工程权衡，公司未披露在达到该效率时的精度保持能力与背隙指标。
Here lies a key unanswered question: the company says cost is about 50% lower than the compared traditional harmonic solution, but does not disclose the specific model compared, procurement volume, precision grade, or test conditions. Joint module cost comparisons depend heavily on the comparison benchmark; harmonic modules of different brands, specifications, and batch sizes vary significantly in price. If the comparison target is imported high-end harmonic modules, the actual price gap between a 50% cost reduction and domestic mature harmonic solutions may be completely different. Similarly, 83% transmission efficiency is within a reasonable range for planetary solutions, but efficiency trades off against precision, backlash, and lifespan; the company did not disclose precision retention capability and backlash metrics at that efficiency.
创始团队背景是理解这笔融资的关键线索。创始人钟雪岗拥有哈尔滨工业大学机械电子工程背景，2019年创办广州傅利叶，三年内主导研发7款康复机器人。此后广州傅利叶被上海傅利叶全资收购，钟雪岗担任傅利叶人形机器人动力板块研发总监，主导20余款执行器、关节模组研发，支持GR1、GR2、GR3等整机产品落地。首席科学家闫磊是哈尔滨工业大学（深圳）副教授、博士生导师，拥有十余年具身力控臂操作算法、灵巧操作前沿研究与产业落地经验。
The founding team background is a key clue to understanding this funding. Founder Zhong Xuegang has a Harbin Institute of Technology mechatronics engineering background; in 2019 he founded Guangzhou Fourier, leading R&D of 7 rehabilitation robots within three years. Guangzhou Fourier was later wholly acquired by Shanghai Fourier, and Zhong Xuegang served as R&D Director of Fourier's humanoid robot power division, leading R&D of over 20 actuators and joint modules, supporting GR1, GR2, GR3 and other complete machine products. Chief Scientist Yan Lei is an associate professor and doctoral supervisor at HIT (Shenzhen), with over a decade of embodied force-controlled arm operation algorithms, dexterous manipulation frontier research, and industrial deployment experience.
这段履历让伊塔动力与傅利叶智能之间的关系变得微妙。傅利叶智能是本轮投资方之一，而伊塔动力创始人曾在其体系内主导动力板块研发。产业资本投资前高管创办的同类硬件公司，在机器人行业并不罕见，但后续的供应链优先级、技术路线区隔和客户冲突，会成为需要持续观察的问题。银河通用的入局同样值得注意。银河通用是具身智能大脑与整机方向的代表性公司，其投资一家关节模组与力控操作平台供应商，逻辑上存在为自身硬件供应链做早期布局的可能。但伊塔动力未披露与银河通用之间是否存在实际采购或联合开发协议。
This resume makes the relationship between Yita Dynamics and Fourier Intelligence subtle. Fourier Intelligence is one of this round's investors, and Yita Dynamics' founder previously led power division R&D within its system. Industrial capital investing in similar hardware companies founded by former executives is not rare in the robotics industry, but subsequent supply chain priority, technology route differentiation, and customer conflicts will become issues requiring continuous observation. Galbot's entry also deserves attention. Galbot is a representative company in embodied intelligence brain and complete machine directions; its investment in a joint module and force-controlled operation platform supplier logically suggests possible early布局 for its own hardware supply chain. But Yita Dynamics did not disclose whether actual procurement or joint development agreements exist with Galbot.
伊塔动力透露，其已与部分具身大脑龙头、人形本体企业及海内外顶尖科研实验室达成战略合作意向，意向金额达数千万元。这是公司目前唯一的商业化进展表述，但战略合作意向与订单之间存在本质区别。意向金额不构成收入，也不代表客户已经完成产品验证、小批量试产或批量采购决策。在硬件供应链中，从意向到订单通常需要经历送样、测试、小批量验证、可靠性评估和商务谈判等多个环节，周期可能长达数月至一年以上。对于一家成立仅五个月的公司来说，数千万元意向金额可以理解为早期市场兴趣的信号，但尚不能作为商业化能力的证明。
Yita Dynamics disclosed it has reached strategic cooperation intentions with some embodied brain leaders, humanoid body enterprises, and top domestic and international research laboratories, with intention amounts reaching tens of millions of yuan. This is currently the company's only commercialization progress statement, but strategic cooperation intentions and orders are fundamentally different. Intention amounts do not constitute revenue, nor do they mean customers have completed product validation, small-batch trial production, or bulk procurement decisions. In hardware supply chains, moving from intention to order typically requires sample delivery, testing, small-batch validation, reliability assessment, and business negotiation—cycles possibly lasting months to over a year. For a company founded only five months ago, tens of millions in intention amounts can be understood as an early market interest signal but cannot yet serve as proof of commercialization capability.
从行业规律看，关节模组客户切换成本高，本体厂商一旦选定供应商，后续更换涉及整机结构、控制算法和供应链管理的重新适配。因此，早期客户往往来自新项目、新机型或对成本敏感的数据采集平台，而非成熟量产机型的直接替换。伊塔动力未披露其意向客户属于哪一类，这直接影响到对其短期商业化速度的判断。光源资本担任本轮融资的孵化方和独家财务顾问，这一角色意味着光源资本不仅参与融资撮合，还在公司早期阶段介入了战略方向、团队搭建或资源对接。
By industry规律, joint module customer switching costs are high; once body manufacturers select a supplier, subsequent replacement involves re-adapting complete machine structure, control algorithms, and supply chain management. Therefore, early customers often come from new projects, new models, or cost-sensitive data collection platforms, rather than direct replacement of mature mass-production models. Yita Dynamics did not disclose which category its intention customers belong to, directly affecting judgment of its short-term commercialization speed. Light Source Capital served as incubation party and exclusive financial advisor for this round—a role meaning it not only participated in financing matchmaking but also intervened in strategic direction, team building, or resource matching in the company's early stage.
伊塔动力面临三重待验证假设。第一重是产品性能数据的独立验证。83%效率、50%成本降幅、2万小时寿命，这些指标需要客户实测或第三方检测来确认，否则难以支撑下游厂商的选型决策。第二重是意向合作的转化效率。数千万元意向金额能否在合理周期内转化为实际订单，是判断其商业模式成立与否的关键。第三重是量产与交付能力。公司成立时间短，核心团队虽有研发与量产经验，但伊塔动力作为独立主体尚未证明其供应链管理、质量控制和批量交付能力。对于硬件公司，量产一致性往往比样机性能更难实现。
Yita Dynamics faces three unverified assumptions. First is independent validation of product performance data—83% efficiency, 50% cost reduction, 20,000-hour lifespan need customer testing or third-party verification to confirm, otherwise difficult to support downstream manufacturers' selection decisions. Second is the conversion efficiency of intention cooperation—whether tens of millions in intention amounts can convert into actual orders within a reasonable cycle is key to judging whether its business model holds. Third is mass production and delivery capability—the company was founded recently; although the core team has R&D and mass production experience, Yita Dynamics as an independent entity has not yet proven its supply chain management, quality control, and batch delivery capabilities. For hardware companies, mass production consistency is often harder to achieve than prototype performance.
从已披露的创始人履历和产品参数看，伊塔动力具备切入关节模组赛道的团队基础和技术方向；从投资方构成看，产业资本的存在为其提供了潜在的供应链入口。但公司目前对外展示的，更多是一个有明确技术路线和早期市场信号的硬件创业项目，而非一个已经通过客户验证的供应商。其真正的分水岭，将出现在首批意向客户是否愿意公开站台、是否产生可确认的采购订单，以及产品能否在批量交付中保持性能一致性。意向金额只是开场，订单才是结论。
From disclosed founder resumes and product parameters, Yita Dynamics has the team foundation and technical direction to enter the joint module track; from investor composition, industrial capital presence provides potential supply chain entry points. But what the company currently displays externally is more a hardware startup project with a clear technical route and early market signals, not a supplier that has passed customer validation. Its true watershed will appear when the first batch of intention customers are willing to publicly endorse, when confirmable purchase orders are generated, and whether products can maintain performance consistency in batch delivery. Intention amounts are only the opening; orders are the conclusion.
