In mid-September 2026, CCTV Finance's Economic Information Broadcast turned its lens to a previously overlooked niche track—robot "skin suits."
Reports show that as the humanoid robot industry accelerates its iteration, robotic skin, a core component for human-machine interaction, has moved beyond the experimental stage: one company secured an order for 13,000 bionic skins in a single month, while another has cumulatively delivered over 40,000 sets of orders, mostly concentrated in the past half year.
The shift in order units from "pieces" to "tens of thousands of pieces" indicates that this track has passed the concept phase. What is even more intriguing is that an order of tens of thousands of pieces has landed on a textile company that produces knitted fabrics.
01. Why a "Coat" Must Be Made to Industrial-Grade Standards
According to the Changshu Converged Media Center, in Changshu, Jiangsu, at the workshop of Jiangsu Jinlong Technology Co., Ltd., the bionic skin customized for robots is embedded with high-strength fibers. Relying on 3D seamless knitting technology, it achieves one-piece molding of the whole machine, and the entire garment is divided into more than 80 functional zones. This "customized robot clothing" uses a deep front-opening zipper to solve the problem of putting on and taking off, and an integrated one-piece design to avoid edge curling and displacement during movement. The joints are also reinforced with anti-stretch treatment, with a stretch rate of over 200%, capable of withstanding millions of bends without significant deformation.
In the product showroom, the humanoid robot wearing pure white knitted clothing keeps the garment fitting the body curve perfectly without wrinkle displacement when raising its arms or turning around. According to Liang Yong, General Manager of the company's business division, the robot's outerwear looks similar to human clothing, but the core requirements are completely different: the way of putting on and taking off is specially designed, and the joint adaptability is also specially designed.
This means it is not a decorative coat, but a component developed according to the logic of industrial parts. The functional requirements are far from those of civilian clothing—according to CCTV Finance, the robot skin needs to simultaneously balance thermal conductivity, anti-static properties, collision buffering, and wear resistance; industrial robots operate all year round, and their exposed joints are most afraid of dust and oil stains; and once they enter shopping malls, nursing homes, or even homes, the sense of alienation brought by metal structures becomes another threshold. According to the Changshu Converged Media Center, before leaving the factory, the products must also pass dozens of industrial-grade tests such as fatigue strength, heat dissipation performance, and electromagnetic compatibility.
02. From "Having Skin" to "Having Tactile Sensation"
If textile technology solves "close-fitting adaptation," then flexible sensing technology makes up for the missing link that allows robots to truly "perceive." Robot skin is thus divided into two layers: the outer layer is responsible for protection, and the inner layer is responsible for sensing, with the latter being where the value lies.
According to CCTV Finance, in Suzhou, Jiangsu, the flexible thin-film pressure sensors produced by Suzhou Leanstar Electronic Technology Co., Ltd. can sense the magnitude of external forces. After being made into electronic skin on robot fingers, the robot can sense the amount of force and control the gripping strength, thereby completing fine operations such as grabbing eggs and driving screws. This product has currently received orders for over a thousand sets. The scenario described by Zhou Zhen, Chairman of the company, is highly visual: robots can now grab eggs and drive screws with their hands. In the future, after entering homes, they will be able to perceive not only touch but also stroking and heating; when shaking hands with them, people can feel that their hands are warm.
In terms of technical routes, two paths are running in parallel. According to CNR News, market insider Shen Rui distinguished them based on the carrier form: one is the thin-film flexible sensing route, mainly piezoresistive, with high sensitivity, suitable for dexterous hands, and the main force of industrialization at this stage; the other is the textile-based fabric sensing route, relying on traditional textile technology, with good extensibility, suitable for large-area coverage of the torso, with large room for cost optimization, but relatively weaker sensing accuracy. In Shen Rui's view, the two routes are not substitutive but complementary—thin films are prioritized for hands, while flexible textile materials are more used for the torso. Notably, the half involving large-area coverage of the torso precisely tests the craft of weaving.
03. Why Old Textile Mills Are the First to Secure Orders
Robot skin requires a flexible substrate that can both carry sensors and fit complex curved surfaces, and "weaving yarn into a close-fitting flexible carrier" is precisely the craft accumulated by the textile industry over decades. The judgment of Sun Zhidong, Senior Researcher at Anxin Fund, is: traditional textile enterprises are deeply engaged in fabric technology and excel at creating flexible carriers for sensors; professional sensor enterprises master core sensing technologies; and the upstream and downstream linkage of the two has rapidly propelled the industry forward. This also illustrates the organizational mode of this round of explosion—it is not about who replaces whom, but about the integration of two industrial chains.
According to the Changshu Converged Media Center, Jiangsu Jinlong Technology is following this path. It has shifted from making clothes for humans to making clothes for robots. The R&D end is involved in the joint development of the whole machine in advance, completing product iteration through 3D modeling and long-term condition field tests, with a development cycle of 6 to 12 months for a single product; orders adopt a progressive mode of "sample—pilot production—mass production," upgrading synchronously with robot hardware. On June 30, 2026, at UBTECH's 2026 Global Launch Event, the company signed an order for 13,000 pieces of robot clothing, covering 10 styles of male and female robot clothing, and became a Tier-1 supplier for UBTECH. Currently, the enterprise has established deep cooperation with multiple leading robot companies such as Zhiyuan, Tuopu Group, and XPeng. Jin Hengjing, Deputy General Manager of the company, said that for traditional textiles to break out of the red ocean of price wars, they must extend to the high end of the industrial chain and seek incremental growth in emerging tracks.
The same capability transfer is even clearer on the equipment side. According to the Securities Times, on August 19, 2026, at the exchange conference on knitted intelligent fabric technology and future applications, Cixing Shares, a global leader in knitting flat machines, jointly signed the "Strategic Cooperation Framework Agreement on Industry-University-Research in the Intelligent Wearable Industry" with material enterprises such as Hangzhou Guorui Superfiber, Kuangda Technology, and Mingxin Xuteng, as well as the College of Textile and Clothing of Qingdao University, the School of Computer and Artificial Intelligence of Wuhan Textile University, and the Jiaxing Research Institute of Southern University of Science and Technology, to collaboratively tackle key technical bottlenecks in intelligent textiles, flexible materials, and intelligent wearables. At the same time, it signed a "Strategic Cooperation Agreement" with Mingxin Xuteng to integrate core resources in material R&D, equipment manufacturing, and process application, building an integrated collaborative system of "material—equipment—process."
Cixing Shares has also independently developed the world's first dedicated computerized flat knitting machine for warp and weft knitting composite weaving for the preparation of knitted flexible electronic skin. It can realize composite weaving of up to 32 different materials. The entire fabric is woven with only one thread, and three-dimensional one-piece molding can be completed without cutting and sewing. The mass production cost is significantly reduced compared to traditional patch-type sensing solutions; although the accuracy of fabric-type electronic skin is not as high as that of polymer thin-film-based sensor arrays, it can achieve full-surface coverage of heterogeneous curved surfaces and is breathable and customizable.
In addition, the company has flexibly introduced more than 10 postdoctoral researchers to jointly develop functional yarns for functions such as pressure detection, temperature measurement, heating, and light emission, and is currently promoting the industrialization of such materials with partners.
From selling equipment to making skin, the craftsmanship in the hands of the old industry is becoming an entry ticket to the new track.
04. After the Order Unit Changes from "Pieces" to "Tens of Thousands"
Changes in numbers illustrate the problem better than any prediction. The unit price of robot skin is not low—the industry usually defines the tactile sensing module adapted to a single finger as "one set" of electronic skin. A humanoid robot often needs to be equipped with 6 to 8 sets, with a total price exceeding CNY 5,000; the cost of a single set of the basic standard version is at the hundreds of CNY level, while for the high-performance version adapted for embodied intelligence model training, the cost of a single set can reach the thousands of CNY level.
The magnitude of orders is also leaping. According to CCTV Finance, the cumulative delivered orders for robot skin from one enterprise have exceeded 40,000 sets, mostly concentrated in the past half year. For a track that has just stepped out of the laboratory, this speed means that demand has shifted from "sample testing" to "batch delivery"—and this is precisely the most difficult watershed to cross in industrialization. The space further ahead has also been estimated: according to predictions by GGII (Gaogong Robot Industry Institute), by 2030, the global demand scale for flexible tactile sensors (electronic skin) in the humanoid robot field will reach 1.525 million square meters, corresponding to a market size of approximately CNY 27.4 billion.
However, the volume increase will not be linear growth. According to CNR News, Shen Rui judges that this process will go through three stages: in the short term of one to two years, it will still be dominated by prototypes; from 2027 to 2029, mass production will accelerate, and applications will expand from hands to the torso; after 2030, it will enter large-scale application, driving cost reduction and efficiency improvement across the entire industrial chain including materials. In other words, the orders of tens of thousands of pieces seen today are just the first stage of the three-stage process.
05. Mass Production Consistency: The Second Answer Sheet for the Old Industry
While orders are exploding, the other real side of the industry also needs to be faced squarely, which precisely points to the next step's growth space.
According to CCTV Finance, in the robot skin production workshop in Jiashan, Zhejiang, basic processes such as sensor film printing have achieved fully automated mechanical operations, but fine links such as sensor fitting and module assembly still highly rely on manual labor at present; another difficulty is standardization. Different robot enterprises have different overall machine shapes and functional requirements, and electronic skins mostly need to be individually customized and developed, making it difficult to release the scale effect for the time being. These two thresholds also mean two spaces—whoever can first make up for automation and establish standards will get the key to scaled production first.
Therefore, the trend of the industry ultimately comes down to two words: delivery and cost reduction. Whether orders can be delivered on time and with quality determines whether an enterprise can truly move from "samples" to "mass production"; whether costs can be brought down determines whether electronic skin can change from an "optional configuration" to a "rigid demand."
Whether robots can perceive collisions, control force, and avoid obstacles in time is directly related to interaction experience and safety—a robot that only performs may not necessarily need skin, but a robot that needs to pick up eggs, tighten screws, and support the elderly must know how much force to use, and this force feedback capability can only be realized through skin. Industry insiders believe that robot electronic skin will show a trend of "rising sales and falling prices."
Therefore, the question to be answered next on this track is already quite clear: it is not about whose launch event is more lively, but about who can hold down both yield and cost at the same time, and who can truly run through the path of "sample—pilot production—mass production."
The process dividends accumulated by the old industry over decades have already cashed in the first order of tens of thousands of pieces; what needs to be cashed in next is the consistency of mass production and long-term delivery capability—and this is precisely the homework that traditional manufacturing is best at.
- End -