Embodied Intelligence Capital Shifts to Robot 'Skin' as Tactile Sensor Funding Rises

By | Machine Frontline

2026's Embodied Intelligence sector, investment trends are shifting.

According to IT Juzi statistics, in October 2026, the embodied intelligence sector saw 33 financing events with a total amount of 10105000000 yuan, while in September the figures were 60 events and 11028000000 yuan, showing a significant downward trend.

However, although the bustling scene of capital chasing body manufacturing and large models has cooled down somewhat, some investors have begun to shift their focus upstream. dexterous hands, tactile sensors, electronic skin these core components that determine a robot's 'physical capabilities' have become segments that continue to attract investment against the trend.

(Image source: official website of Juqiao Industry)

On October 29, Shandong-based flexible fabric tactile sensing company Juqiao Industry announced the completion of its Series A financing round of 100 million yuan, led by Ant Group, with existing shareholders Zero Yi Capital, Weihai Gangrong, and Qingdao Hexin participating as co-investors. This marks Ant Group's second move in the tactile sector; two months ago, it had led the investment in the visual-tactile route of Dai Meng Robotics.

It is worth noting that the tactile sector features highly fragmented technical routes; a company established less than three years ago securing on the order of 100 million yuan in financing—does this signal an industry turning point or merely a temporary capital positioning?

01 The financing logic of embodied intelligence has changed

The financing logic for embodied intelligence seems to be shifting from 'building a smart body' to 'enabling the body to perceive the world'.

After all, vision can be fed by internet videos, language by text, but tactile data is almost impossible to obtain from existing corpora. Physical quantities such as tangential force and material texture are difficult for simulation engines to generate with high fidelity; they must be collected frame by frame through real-world operation.

Industry data shows that in terms of price, tactile data costs about twice as much as visual data. Video data such as vision may cost a few hundred yuan per hour, while tactile data costs several hundred yuan or more per hour.

This difference is directly reflected in financing figures. In the first seven months of 2026, nearly 20 financing rounds occurred in China's humanoid robot sensor sector, totaling over 3100000000 yuan, with about 1560000000 yuan alone absorbed in July. To some extent, this also means that whoever can provide an efficient, low-cost, and scalable tactile data acquisition entry point next will control the current 'data bottleneck”.

However, China's tactile track currently does not follow a single unified technical route. The Hall magnetic sensing solution, represented by Parini, leans toward high-precision multi-dimensional force detection at the tips of dexterous fingers. The visual-tactile route, represented by Dai Meng Robotics, has higher resolution; its tactile resolution can reach tens of thousands of sensing points per square centimeter, outputting image data that naturally adapts to end-to-end networks.

There is also a fabric-based route, represented by Juqiao Industry. Fabric-based approach Its technical logic involves decomposing conductive and piezoresistive structures down to the fiber level, weaving sensing networks directly from functional fibers. When external force touches the fabric, the conductive pathways within the fibers deform, allowing the acquisition system to read signals and algorithms to convert them into tactile data.

These three approaches are not mutually substituting; rather, they correspond to different tasks respectively. For instance, fingertips and dexterous hands require high resolution, multi-dimensional force sensing, and fast response, where visual-tactile and Hall magnetic sensing solutions currently hold an advantage. Large curved surfaces such as the torso, robotic arms, and soles of feet need extensive coverage that is bendable, breathable, and cost-controllable, making the fabric-based approach seem easier to scale.

Therefore, it is difficult for any single approach to dominate in the short term; the industry is more likely to evolve toward layered hybrid solutions based on body parts. Consequently, capital investment logic does not bet on one technology ultimately winning. Ant's two investments precisely reflect this differentiation: Dai Meng Robotics represents the high-precision direction of visual-tactile sensing, while Juqiao Industry represents the large-area coverage capability of the fabric-based approach. This also explains why, against the backdrop of cooling financing in embodied intelligence, a sum on the order of 100 million yuan still falls on 'a piece of cloth'.

02 'Counterintuitive' Technical Path

Capital investing in both approaches simultaneously does not mean that both paths have been fully validated. To determine whether this on the order of 100 million yuan financing signals an industry inflection point or merely a phased capital positioning, we must return to Juqiao Industry itself.

The company was founded at the end of 2023. Founder Zhang Jingqi graduated from the Department of Electronic Engineering at York University in the UK and previously served as CTO of a Xiaomi ecosystem company, holding multiple related invention patents. The core team mainly comes from the Chinese Academy of Sciences and major smart hardware manufacturers, possessing deep technical accumulation as well as experience in defining and mass-producing smart hardware products.

The company's technical path exhibits a notably 'counterintuitive' character. In traditional piezoresistive films, the structure involves sequentially stacking flexible substrates, pressure-sensing layers, and conductive layers. However, Juqiao Industry chooses to use nylon as the fiber core, fabricating piezoresistive materials and electrodes as independent functional fibers respectively.

The advantage of this solution lies in that the resulting sensing fabric can be cut, folded, and rapidly redesigned, enabling it to conform to complex curved surfaces of robots. This innovation successfully propelled fabric-based electronic skin toward true mass production. In May 2025, the company disclosed a roll-to-roll production line with an annual capacity of 200,000 units and a yield rate of 95%; in July, Weihai Lingang District reported that orders were already scheduled through October, with first-year output upon production nearing 10,000,000 yuan; in September, the 1,000th set of humanoid robot fabric electronic skin was completed, and the company stated that with less than half a year since product launch, nearly 100 domestic and international customers had placed orders and repurchased, including but not limited to Unitree, Songyan Dynamics, ZeroCube, and others.

However, the shortcomings of the fabric route are also evident: its single-point precision is inferior to visual-tactile solutions, making it more suitable for large-area coverage scenarios such as the torso, soles, and mechanical arms rather than high-precision fingertip operations. Therefore, the company is also extending into fine-grained interaction by launching Tactile Skin 2.0, which further expands perception capabilities for force direction, sliding trends, and tangential forces through adjustments to weaving structure combined with model training, but whether it can meet the high-end demands of dexterous hands in terms of precision remains to be verified.

Deeper challenges lie in the business model. It is reported that Juqiao Industry's current product matrix spans three tracks, among which robot electronic skin serves as a valuation anchor, establishing the company's technical benchmark status; however, at this stage, business is highly dependent on a few leading customers such as Unitree, and order sustainability remains to be observed.

Smart health and wellness serves as a cash-flow business, with orders exceeding 10 million yuan in 2024, and implementations at renowned institutions such as Taikang Home, SDIC Health, and China Resources Runjia; conversely, the intelligent cabin for automobiles acts as a growth lever to support performance.

Therefore, the meaning of this on the order of 100 million yuan financing is more accurately that it does not signify that capital has cast a 'confirmation vote' for Juqiao Industry's fabric route, but rather positions itself early during the window period when tactile technology routes have not yet converged and data entry points are not yet settled, securing a differentiated player with mass production delivery capabilities.

What it truly tests is whether the company can convert its early-moved process advantages into stable delivery, standardized modules, and tactile data usable by models; if this step succeeds, the fabric route has the potential to become the infrastructure for whole-body robot touch, whereas if it fails, the capital window may be rapidly compressed by later entrants.

03 Who Can Seize the High Ground of Humanoid Robot Perception?

The reason why Juqiao's financing has attracted industry attention actually reflects changes occurring in robot skin.

In the past, robots wearing clothes or being covered with soft skins were more often understood as issues of appearance and protection. Now, this outer covering simultaneously carries four functions: safety, affinity, scene adaptation, and tactile perception.

In near-human scenarios such as shopping malls, health care, and homes, whether users encounter hard metal or flexible covering directly determines the experience of human-robot coexistence; in industrial scenarios, heat resistance, corrosion resistance, dust-proofing, and static dissipation are essential requirements...

This has also led to particularly significant growth in the sector. According to industry forecasts, demand for human-shaped robot electronic skin in China will reach 1.525 million square meters by 2030, corresponding to an overall market size of approximately 27,400,000,000 yuan. Traditional textile enterprises, specialized sensor companies, and material firms are accelerating their entry into the market, with obvious linkage effects between upstream and downstream parts of the industry chain.

However, behind the heat, the industry development is still in its early stage. The lack of standardization is one of the most prominent pain points. Robots of different brands and models have varying structures; electronic skin mostly requires custom development for each unit, so molds and patterns cannot be reused in bulk, leading to high costs and long cycles.

High-end electronic skin modules are constrained by packaging processes, sensing chips, and material costs, making the full set price still difficult for cost-effective whole-machine manufacturers to bear. In addition, flexible sensing products may experience signal drift, reduced sensitivity, shortened service life, and other issues under conditions of long-term high-frequency bending, complex industrial dust, and temperature variations; product stability and consistency need improvement. For robot skin to truly transition from an optional feature to a must-have, it must cross three thresholds: consistency, reliability, and cost.

In summary, for a machine to enter human spaces, what is perceived first is often not algorithms or parameters, but whether it feels hard or soft, cold or warm when touching a person; the value of tactile sensing lies precisely in this.

However, the maturity of this track will not be marked by a financing round or a new product, but by skin becoming an unobtrusive yet indispensable ordinary component. At that time, the focus of discussion will no longer be on which technical route is more advanced, but on who can stably supply under real operating conditions, control costs, and cooperate with whole-machine iteration. The capital betting today is precisely on this certainty. As for which route ultimately wins, the answer may only come after robots truly enter daily life.

(The author of this article is Machine Frontline; published by Ti Media with authorization)