Hundred-billion unicorn, taking a chance for humanoid robots
「Key Takeaway」
The technical routes for dexterous hands have not converged yet. Whoever can achieve large - scale delivery first will secure the core ecological niche for robot "hands".
The hand is the most delicate part of the human body. The two hands have 54 bones, accounting for a quarter of the total number of bones in the body. When humanoid robots attempt to replicate this dexterity, the "hand" has become one of the most difficult hurdles to overcome.
Even Tesla, a powerful company, has stumbled on this hand. The mass - production plan for Optimus has been postponed again and again. In the third quarter of 2025, Elon Musk confirmed that the design finalization of the third - generation product would be postponed to the first quarter of 2026. What's holding it back is precisely that hand. According to the technology media The Information, in the express sorting training, the service life of the dexterous hand is only 6 weeks, while the cost of a single hand exceeds $6000.
Short service life, high cost, and insufficient degrees of freedom have turned countless "physically strong" humanoid robots into "semi - finished products" that have a good frame but can't do the work. This has also made the dexterous hand an important lever to pry the embodied intelligence industry.
In February this year, a domestic company called "Lingxin Qiaoshou" completed a Series B financing of 1.5 billion yuan, with a valuation exceeding 10 billion yuan, becoming the company with the highest valuation in the dexterous hand field.
The fact that Lingxin Qiaoshou is being pursued by capital actually means that everyone is betting on who in the dexterous hand track can be the first to move from "laboratory customization" to industrial standard products.
Whether Lingxin Qiaoshou can become a real "hidden champion" depends on whether the delivery data in 2026 can be fulfilled and whether the humanoid robot industry itself can achieve the expected volume.
Capital Chases "A Pair of Hands"
The dexterous hand is becoming a very crowded sub - track in the embodied intelligence industry chain.
According to incomplete statistics from Xinchacha, currently there are more than 80 companies globally engaged in the research and development of dexterous hands, among which Chinese companies account for more than half of the share.
The players in this track can be roughly divided into three categories: independent integrators focusing on dexterous hand research and development, such as Lingxin Qiaoshou, Lingqiao Intelligence, Sharpa, etc.; humanoid robot body companies, such as Tesla and Unitree Technology; and manufacturing enterprises extending downstream from core components, such as Pacini Perception.
Capital is voting for this track with real money. Lingxin Qiaoshou, a leading player in the dexterous hand track that completed a Series B financing in February 2026, is only 7 months away from its previous round of financing, and now the company's valuation has increased tenfold.
Zhou Yong is the founder of Lingxin Qiaoshou and graduated from Huazhong University of Science and Technology. Before founding Lingxin Qiaoshou, he had two entrepreneurial experiences. The first was in the game and community business, and the second was in intelligent outdoor unmanned vehicles.
The motivation to engage in cutting - edge technology comes from his childhood. Because Doraemon has a magic pocket that can take out whatever he wants, this made Zhou Yong dream of replicating the human tool for creating all things - the hand - with machinery.
Public information shows that the pre - preparations for Lingxin Qiaoshou started in 2019. By 2023, Lingxin Qiaoshou became an independent company, and in the following year, it released the first - generation product of Linker Hand. In early 2025, it released the durable models (T10, T20), high - end model (Ultra), and entry - level model (O7) of the same series, and at the same time prepared for financing to support mass production.
In addition to Zhou Yong, the company's co - founder team also includes Su Yang, the chief AI architect; Zhang Yanbai, the person in charge of marketing; Zuo Jiaping, the former senior vice - president of Zhipu; and Jia Xiaoyou, who is in charge of product design.
Among them, Jia Xiaoyou is a young post - 2000s. He graduated from undergraduate in 2025 and was the chairman of Zhongda Hengji (Beijing) Digital Economy Technology Development Co., Ltd., which is also an early shareholder of Lingxin Qiaoshou. His development in the dexterous hand field has made Jia Xiaoyou included in the "Forbes China 30 Under 30" elite list in 2025.
From April 2025 to the present, Lingxin Qiaoshou has raised funds 6 times in total, with a cumulative financing amount of billions of yuan. The investment companies include Sequoia, Ant Group, Zhejiang Innovation Investment, etc.
What investment institutions value first is the track. In addition to the "fear of missing out" brought by the prosperity of the humanoid robot industry, it is more about the great significance of the "hand" to humanoid robots: as the end - effector of the robot, the state of the dexterous hand determines the final production and service effect of the humanoid robot.
The importance of the dexterous hand can be seen from the cost proportion.
Currently, the cost of a dexterous hand is about 20% of the whole machine. If only the motion system is considered, the cost of the hand even exceeds 30%. As the requirement for the fine operation of the hand increases in the future, this proportion is likely to further increase.
Elon Musk has emphasized the extreme importance of the dexterous hand on multiple occasions. He frankly said in an earnings call: "The fine operation of the hand is the most difficult part of the whole design. It is a very difficult thing to create a hand as dexterous and capable as a human hand."
In this competition, whoever can achieve large - scale delivery first will secure the core ecological niche for robot "hands". This is also the key for Lingxin Qiaoshou to be favored by capital.
Breaking Through in Mass Production with a "All - in" Strategy
According to the understanding of "Baobian", currently Lingxin Qiaoshou receives orders of about a thousand units per month. The products are mainly sold to universities, whole - machine manufacturers, and factories.
Lingxin Qiaoshou has about 300 to 400 employees. From hardware transmission to AI algorithms, hundreds of people are focused on the dexterous hand.
But currently the technology of dexterous hands has not converged to an optimal route. Instead, it presents a situation where tendon - driven, direct - drive, and linkage - drive are in parallel. Tendon - driven transmission has good flexibility and a compact structure, but "creep" will occur after long - term use, resulting in a decrease in hand control accuracy; direct - drive transmission has high accuracy and strong rigidity, but is limited by large volume and poor heat dissipation; linkage - drive has high stiffness and low cost, but lacks flexibility and is difficult to adapt to complex grasping scenarios.
Different from other manufacturers, Lingxin Qiaoshou invests resources in all possible technologies, choosing to "not bet on one route, but embrace all", covering all the technical routes on the market.
This "no - betting - on - one - route" strategy is clearly reflected in the product matrix: Linker Hand L20 uses a linkage - drive solution, targeting scientific research and industrial fine assembly; L30 uses tendon - drive; L6 uses a direct - drive solution for high - load industrial scenarios. The motivation for this strategy is that no matter where the technology converges, Lingxin Qiaoshou will not fall behind.
But this also means the dispersion of R & D resources. If the technical route fails to converge for a long time, the multi - line layout may face the pressure of technological iteration and production capacity integration; even if the technical route finally converges, Lingxin Qiaoshou may find it difficult to significantly outperform other competitors because it has spent a lot of resources on other routes.
The full coverage of technical routes has brought the result of product diversification. Currently, Lingxin Qiaoshou has a total of 7 series and more than 10 sub - models, with prices ranging from 3999 yuan to 100,000 yuan.
The entry - level series includes the O6 (6666 yuan, linkage - drive, 6 degrees of freedom, the lightest in the world at 370g) released in August 2025, and the O7 (8800 yuan, linkage - drive, 7 degrees of freedom) released at the Industrial Expo in March 2025.
The industrial main - force series includes the L6 (19,900 yuan, 6 degrees of freedom, linkage - drive), L20 (49,900 yuan, linkage - drive, 20 degrees of freedom), L30 (100,000 yuan, tendon - drive, 25 degrees of freedom), etc.
But no matter how the technical route converges, the final form of the robotic hand must be at least as flexible as a human hand. A study published in "Nature Communications" by the University of Science and Technology of China in January 2025 clearly pointed out that a human hand has 23 degrees of freedom (24 - 27 if wrist movement is included). Currently, companies that have achieved this degree of freedom in dexterous hands include Lingxin Qiaoshou, Tesla, Shadow Robot, Yihai Yuanshi, Lingqiao Intelligence, Xingji Guangnian, etc.
Among these manufacturers, Lingxin Qiaoshou has clearly achieved mass production, and the most popular is the entry - level O6 series. The price of this dexterous hand is about 6666 yuan, and the O6 lite is even sold at only 3999 yuan, far lower than the price of Tesla's dexterous hand.
Lingxin Qiaoshou can lower the price of dexterous hands. "First of all, it comes from the extremely low industrial production cost in China." Su Yang once mentioned in an interview with the media: "Compared with Europe and the United States, the prices of industrial parts in the Yangtze River Delta and the Pearl River Delta are very cheap. Many people don't know that it is also very cheap around Beijing."
Another reason is self - research. Lingxin Qiaoshou independently develops 100% of its core components (motors, reducers, electric cylinders, sensors), and uses engineering plastics to replace traditional metal reducers, reducing the weight by 30% - 40% and the material cost from hundreds of yuan to dozens of yuan.
Realizing mass production first also means a data advantage. As the "only company in the world with a monthly production of a thousand high - degree - of - freedom dexterous hands", Lingxin Qiaoshou can obtain a large amount of training data earlier, which is the scarcest "fuel" for the development of robot intelligence.
Is the "Ceiling" of the Dexterous Hand in Someone Else's Hands?
Lingxin Qiaoshou is leading the way, but the road ahead is not smooth. The dexterous hand track faces a fundamental paradox: as the "hand" of the robot, its upper limit depends not only on how dexterous it is, but also on how many whole - machine humanoid robots are mass - produced.
If humanoid robots cannot form sufficient commercial value in the foreseeable future, the dexterous hand can only "quench its thirst by looking at plums". In 2026, Lingxin Qiaoshou set a delivery target of 50,000 to 100,000 units. To some extent, this is "betting" on the industry to achieve the expected volume.
In addition, the training of the hand cannot be completed in isolation in essence. The tactile feedback and force - control information contained in the real - machine operation data must be collected relying on the complete coordination of the body - arm - hand.
Academician Yao Qizhi pointed out at the 2025 Global Developers' Pioneer Conference: "The motion ability and operation ability of existing robots are still two separate systems. In the future, unified whole - body control and fine - hand operation planning are needed to support multi - step general skills."
As a component supplier, Lingxin Qiaoshou cannot control the whole - machine motion control (the "cerebellum" of the robot) on the one hand, and on the other hand, directly collecting data in various scenarios is extremely costly. Its technological iteration is objectively restricted by the maturity of the humanoid robot ecosystem.
However, the reality is that the global stock of humanoid robots is extremely limited. The cost of collecting real - machine data is as high as 20 yuan per piece. Even if a hundred robots are deployed, the daily output is only 80,000 - 100,000 pieces.
Moreover, the real - life environment that robots will face in the future will only be more complex. The models trained with data collected through remote operation in the laboratory often "feel helpless" in real scenarios.
Currently, Lingxin Qiaoshou is trying to solve this problem through the Open TeleDex open remote - operation system and the LinkerSkillNet multi - modal data collection system, but the gap between the "hand" and the "brain" is still a chasm that the entire industry needs to cross.
He Xiaopeng, the CEO of XPeng Motors, once said bluntly that the current development stage of humanoid robots is only equivalent to the initial stage of L2 in autonomous driving, and it still needs to cross the "threshold of dozens of times of ability and difficulty" to reach the initial stage of L3 with commercial value.
Before crossing the threshold, Lingxin Qiaoshou still needs to continue this expensive money - burning game.
This article is from the WeChat official account “Baobian” (ID: baobiannews), author: Zhang Jingwei, published by 36Kr with authorization.