日前，国内具身智能企业卫澜深海获得了由深创投、元禾原点、国策资本等在内的新一轮投资，主要用于训练水下具身智能模型、建设训练中心及拓展海洋油气场景。其实具身智能行业获得融资是一件很平常的事情，但诡异的地方在于卫澜深海获得融资的频次实在是太高了。这家公司成立于2021年9月，此前一直不温不火，但是从去年12月开始，卫澜深海在9个月时间内获得了5轮融资，融资速度和节奏堪称劳模标杆。
Recently, domestic embodied intelligence company Weilan Deepsea obtained a new round of investment from Shenzhen Capital Group, Oriza Origin, Guoce Capital, and others, mainly for training underwater embodied intelligence models, building training centers, and expanding marine oil and gas scenarios. Actually, funding in embodied intelligence is a very ordinary thing, but the strange part is that Weilan Deepsea's funding frequency is really too high. Founded in September 2021, the company had been lukewarm before, but starting last December, Weilan Deepsea obtained 5 funding rounds in 9 months, with a pace worthy of a model worker benchmark.
为什么这样一家在过去五年内不起眼的具身智能公司，却能在最近九个月获得资本的频繁关注。一方面在于其水下具身智能的独特定位，另一方面也是如今具身智能行业同质化严重的反衬。在所有具身智能公司里，卫澜深海是一家比较另类的公司，它的官网刚刚上线，像产品介绍、未来规划等常规资料，一概没有，而且从2021年9月截至2025年12月的四年半时间内，卫澜深海都与资本没有交集，在具身智能最火热的2024、2025年，它没有获得任何融资。
Why could such an unremarkable embodied intelligence company over the past five years get frequent capital attention in the last nine months. On one hand, its unique positioning in underwater embodied intelligence; on the other, it contrasts sharply with today's severe industry homogenization. Among all embodied intelligence companies, Weilan Deepsea is a relatively alternative one—its official website just launched, with no conventional materials like product introductions or future plans, and from September 2021 to December 2025, four and a half years, Weilan Deepsea had no capital intersection—during the hottest 2024 and 2025 for embodied intelligence, it obtained no funding at all.
但就是这样一家默默无闻的科技创业企业，在2026年却像开了挂一样。2025年12月，卫澜深海先后两次获得由英诺天使基金和奇绩创坛两家投资机构天使轮、Pre-A轮投资。今年7月，获得元禾原点、国策资本、英诺天使、深创投的A轮投资。10月获得深创投、元禾原点等资本在内的接连两轮投资。为什么这样坐冷板凳具身智能企业突然间在2026年崛起逆袭了呢。原因不在于技术，而是定位。
But such an obscure tech startup seemed to have cheats enabled in 2026. In December 2025, Weilan Deepsea successively obtained angel and Pre-A rounds from Inno Angel Fund and MiraclePlus. In July this year, it obtained Series A investment from Oriza Origin, Guoce Capital, Inno Angel, and Shenzhen Capital Group. In October, it obtained two consecutive rounds including from Shenzhen Capital Group and Oriza Origin. Why did this bench-warming embodied intelligence company suddenly rise and counterattack in 2026. The reason is not technology but positioning.
现在的具身智能企业看似正处于烈火烹油的繁荣时期，但很多人会发现，具身智能最大的问题在于同质化太严重了。只要打开具身智能机器人公司的官网，在产品介绍时你会发现，这些产品的定位基本就三个方向，要么进入家庭代替保姆育儿嫂，照顾老人和孩子，要么进入工厂，在流水线上代替工人搬运物料打螺丝，要么去博物馆科技馆当无人讲解员。可以说，不论是宇树科技、智元机器人，还是优必选等，他们的具身智能产品应用的领域基本就这几样。而国内智源研究院给出的数据显示，目前国内做具身智能的企业数量达到230家，上百家企业的上千种具身智能产品，去争夺家用或是工厂领域，显然肉少狼多。按照科技产品的马太效应，一旦具身智能在技术和落地场景上迎来重大突破，那些同宇树、智元、优必选做不出差异化的企业，绝大部分都会被市场淘汰掉。所以对于资本来讲，具身智能是一个大方向，是未来的核心产业，但为了避免同质化造成的风险，他们必须找到差异化的蓝海赛道。
Today's embodied intelligence companies appear to be in a fiercely prosperous period, but many will notice the biggest problem is severe homogenization. Just open any embodied intelligence robot company's website and in product introductions you will find these products basically positioned in three directions—either entering homes to replace nannies and childcare workers, caring for elderly and children; or entering factories to replace workers on production lines carrying materials and tightening screws; or going to museums and science museums as unmanned guides. It can be said that whether Unitree, AgiBot, or UBTECH, their embodied intelligence products basically apply in these areas. Data from the Beijing Academy of Artificial Intelligence shows domestic embodied intelligence companies number 230, with hundreds of companies' thousands of embodied intelligence products competing for home or factory sectors—clearly too little meat for too many wolves. According to the Matthew effect of tech products, once embodied intelligence achieves major breakthroughs in technology and deployment scenarios, companies that cannot differentiate from Unitree, AgiBot, and UBTECH will mostly be eliminated by the market. So for capital, embodied intelligence is a major direction and future core industry, but to avoid homogenization risks, they must find differentiated blue ocean tracks.
而在所有的具身智能项目中，卫澜深海显得有些特立独行，他的产品不服务家庭也不服务于工厂生产线，而是定位在海洋里。海洋有很多工作场景，如海底电缆光纤的巡检，海洋里面石油钻井平台的维护保养等，传统上这些都需要依靠拖船将人和设备拖到附近进行操作。卫澜深海旗下的超级蛙人号海洋具身智能机器人由水下机器人通用平台和水下原位供电基站两部分组成，前者能够代替人工进行水下作业，后者能够为机器人充电补能。据卫澜深海数据，该机器定位精度误差不超过1米，目标识别准确率超过95%，搭载双6自由度机械臂，负载10公斤以上，单次续航48小时。
Among all embodied intelligence projects, Weilan Deepsea appears rather maverick—its products serve neither homes nor factory production lines but are positioned in the ocean. The ocean has many work scenarios, such as submarine cable and fiber optic inspection, offshore oil drilling platform maintenance—traditionally requiring tugboats to drag people and equipment nearby for operations. Weilan Deepsea's Super Frogman ocean embodied intelligence robot consists of two parts—an underwater robot general platform and an underwater in-situ power supply base station. The former can replace manual underwater operations; the latter can recharge the robot. According to Weilan Deepsea data, the machine's positioning accuracy error does not exceed 1 meter, target recognition accuracy exceeds 95%, equipped with dual 6-DOF robotic arms, payload over 10 kg, single endurance 48 hours.
这种独特的应用定位，让卫澜深海与目前国内大多数的具身智能企业实现了错位竞争。一是它避开了智元、宇树这些行业第一梯队在家庭看护、工厂应用这两大领域的竞争，转向海洋应用领域，避免了与这些巨头的短兵相接。并且海洋环境对具身智能有商业需求，例如传统油田维护，如果采用海洋具身机器人，能够帮海洋石油客户每年节省一亿以上的船舶租金费用。二是它确立了自己的技术优势，像智元、宇树这些头部厂商，他们的产品实力很强，可以复制到很多领域，如产线打螺丝的机器人经过升级后可以到物流去搬运，具备一定的通用性。但因水文数据、探测设备等问题，对于智元、宇树等传统具身智能企业来说，海洋具身智能在硬件和软件层面都是一个巨大的壁垒。
This unique application positioning lets Weilan Deepsea achieve differentiated competition with most domestic embodied intelligence companies. First, it avoids competition with industry first-tier players like AgiBot and Unitree in home care and factory applications, turning to ocean applications, avoiding direct confrontation with these giants. And the ocean environment has commercial demand for embodied intelligence—for example, traditional oilfield maintenance using ocean embodied robots can help offshore oil customers save over 100 million yuan annually in vessel rental costs. Second, it establishes its own technical advantages. Head players like AgiBot and Unitree have strong product capabilities replicable to many fields—for example, production line screw-tightening robots can be upgraded to logistics handling, possessing certain generality. But due to hydrological data, detection equipment, and other issues, for traditional embodied intelligence companies like AgiBot and Unitree, ocean embodied intelligence is a huge barrier at both hardware and software levels.
对于大众来看，具身智能整个行业正处于烈火烹油的繁荣时期，不论是亦庄的机器人大会，春晚上具身智能的表演，还是宇树科技上市IPO，整个具身智能产业正在加速进化。但对于产业和资本来说，他们则是与普通大众不一样的体感，其中最有代表性的就是朱啸虎对具身智能产业的点评，中国现在有差不多200家公司，绝大部分会死，商业化太小了，我不喜欢机器人，那些只能跳舞的机器人，200家人形机器人公司差异太小。从朱啸虎的观点来看，具身智能不是没有商业化，而是这么多企业进入具身智能，但是这些企业的产品设计和定位高度相同。
To the public, the entire embodied intelligence industry is in a fiercely prosperous period—whether it is the Yizhuang robot conference, Spring Festival Gala embodied intelligence performances, or Unitree's IPO, the entire embodied intelligence industry is accelerating evolution. But for industry and capital, their experience differs from the general public, most representatively Zhu Xiaohu's commentary on the embodied intelligence industry: China now has about 200 companies, most will die, commercialization is too small... I don't like robots, those robots that can only dance, 200 humanoid robot companies are too similar. From Zhu Xiaohu's view, embodied intelligence is not lacking commercialization, but so many companies entering embodied intelligence have highly similar product designs and positioning.
而水下具身智能为投资机构和创业者提供了新的思路，当行业在家政服务、工厂打螺丝这些场景中搏杀时，海洋则是新通道。其实，在具身智能同质化浪潮中受益的不只是卫澜深海这一家公司。工至海洋一年内融资三次，其中2026年8月、9月相继获得两次融资，融资金额达到亿级。世航智能在2025年12月、2026年6月，分别获得数亿元的A轮和Pre-A+轮投资、超10亿元的A轮系列投资，作为连续跟投5次的投资人，朱啸虎甚至表示，SpaceX重新定义了太空，世航智能正在用具身智能重新定义海洋。深海智人则在2026年8月获得了超5亿元的A轮投资，该笔投资由中保投资、中电科投资、道达尔能源等国家级资本和能源产业资本组成，被称为史上最强A轮。
Underwater embodied intelligence provides new ideas for investment institutions and entrepreneurs—while the industry battles in domestic services and factory screw-tightening scenarios, the ocean is a new channel. Actually, Weilan Deepsea is not the only company benefiting from the embodied intelligence homogenization wave. Gongzhi Ocean raised three times in one year, including two rounds in August and September 2026, with funding reaching hundreds of millions. Shihang Intelligent obtained hundreds of millions in Series A and Pre-A+ investment in December 2025 and June 2026 respectively, and over 1 billion yuan in Series A rounds; as an investor following five consecutive rounds, Zhu Xiaohu even said SpaceX redefined space, and Shihang Intelligent is redefining the ocean with embodied intelligence. Shenhai Zhiren obtained over 500 million yuan in Series A investment in August 2026, composed of national-level capital and energy industry capital including China Insurance Investment, CETC Investment, and TotalEnergies, called the strongest Series A in history.
从上述我们可以看出，在具身智能这个赛道上，资本市场正在规避同质化竞争引发的投资风险，将资金投入冷门的水下具身智能领域。可以说，卫澜深海等水下具身智能创业企业的投资潮，既得益于自身的商业价值，但更多的是取决于时代的转变，具身智能在工厂打螺丝、家庭护理上的同质化布局，让水下这个冷门不能再冷门的领域成为具身智能的新风口。从卫澜深海到工至海洋、世航智能、深海智人，水下具身智能赛道正在批量制造融资纪录。这并非偶然，而是资本在具身智能同质化困局中，用真金白银投票选出的蓝海样本。
From the above we can see that in the embodied intelligence track, capital markets are avoiding investment risks caused by homogenized competition, putting money into the niche underwater embodied intelligence field. It can be said that the investment wave in underwater embodied intelligence startups like Weilan Deepsea benefits both from their own commercial value and more from the era's shift—embodied intelligence's homogenized布局 in factory screw-tightening and home care has made the underwater field, once impossibly niche, a new outlet for embodied intelligence. From Weilan Deepsea to Gongzhi Ocean, Shihang Intelligent, and Shenhai Zhiren, the underwater embodied intelligence track is mass-producing funding records. This is not accidental but a blue ocean sample capital voted for with real money in the embodied intelligence homogenization dilemma.
