Ant Group has led a ¥100 million Series A in Juqiao Industrial, a Weihai-based maker of fabric electronic skin. Existing shareholders Lingyi Venture Capital, Weihai Gangrong, and Qingdao Hexin also participated. The company plans to expand production and delivery, develop tactile and multimodal technology, and pursue robotics, home, and healthcare applications.
TMTPost reported the financing on October 29, 2026, citing financial adviser Duowei Capital. Gasgoo published a fuller account on October 31. Neither report provides a separate closing date.
The investment backs a physical component with consequences for AI training and robot control. Juqiao weaves sensing functions into textiles, aiming to give robotic surfaces a sense of contact and let human-worn gloves collect demonstrations that include touch.
Why Robots Need Touch Alongside Vision
A camera can show a robot where a cup sits and whether its fingers are approaching the rim. Contact introduces another set of problems: how tightly to squeeze, whether the cup is sliding, and whether one finger is carrying too much of the load.
Three signals help explain the difference:
- Normal force acts perpendicular to a contact surface, such as a fingertip pressing into a cup.
- Shear force acts along the surface, such as a held object pulling downward against the fingers.
- Slip tendency indicates that a contact may be losing its grip, giving a controller a reason to adjust.
Vision can estimate some of these conditions from motion and deformation. It cannot directly measure the mechanical load at a hidden contact point. A camera might see that a cup is deforming, but a tactile sensor can provide a local signal before the deformation becomes visually obvious.
Gasgoo describes cup handling and part insertion as examples of tasks that benefit from tactile feedback. Juqiao says its latest Skin 2.0 platform detects force direction, shear, and slip tendency. Those capabilities are company-reported.
The practical value is a richer feedback loop: look to approach an object, then use contact signals to adjust the interaction. That still requires a controller capable of interpreting those signals and acting on them quickly.
How Juqiao Turns Fabric Into a Sensor
Juqiao’s approach assigns conductivity, pressure resistance, and load-bearing functions to separate fibers, then weaves them into a sensing network. Contact deforms the local structure and changes conductive pathways. Electronics read those changes, while calibration and algorithms translate them into tactile values.
The sensing structure is therefore part of the textile itself. That creates a route to flexible coverage across gloves and curved robot surfaces, although flexibility alone says little about durability or measurement quality.
Gasgoo reports the following Juqiao specifications:
| Company-Reported Specification | What It Means |
|---|---|
| Hundreds of sensing points per hand | Contact can be measured at multiple locations across the glove |
| Minimum sensing-point spacing of 2 mm | The closest points are separated by two millimeters; this is not a claim of uniform spacing |
| Force detection as low as 0.01 newtons | A low-force detection threshold, not a statement of accuracy across the operating range |
| Multilayer sensors below 0.8 mm thick | A thin structure for layered sensing |
| Single-layer products below 0.3 mm thick | A separate thickness claim for single-layer flexible products |
| Approximately ±5% consistency when mapping readings to mechanical values | A consistency claim whose test conditions are not disclosed in the report |
The distinction between detection threshold and accuracy matters. Detecting a very small force does not tell a buyer how accurately the sensor measures larger forces, how quickly it responds, or how readings change after repeated bending.
These specifications and the commercial figures below come from company statements carried by the reports, without independent testing or auditing. The published accounts provide limited measurement conditions, so they support a description of Juqiao’s claimed capabilities, not a verified performance ranking.
Human-Worn Gloves Are Also Robot Training Tools
Juqiao’s tactile gloves have two different roles in the embodied AI story. They can support motion-capture and teleoperation setups, and they can record what people feel while performing tasks.
According to Gasgoo’s account, a person wearing the gloves generates records of contact areas, force changes, and movement details alongside first-person video. Juqiao is also developing a collection system supported by software development kits, visualization software, and training tools.
That pairing could make demonstrations more informative. A video might show someone lifting an object successfully. Synchronized touch data could additionally show when contact began and how grip force changed during the lift.
This is a training-data argument, separate from putting sensors on a working robot. Human demonstrations still need to be translated into actions suited to the robot’s hand geometry, actuators, and control system. A large collection of recordings does not automatically solve that transfer problem.
I’d treat Juqiao’s glove shipments as evidence of its reported data-collection footprint. Counting them as tactile-equipped robots would confuse two different products and deployment stages.
What the Production Numbers Actually Show
Juqiao says it has served more than 400 clients, including over 200 in embodied intelligence, and shipped nearly 10,000 tactile gloves. It also reports production yield above 95% on its Weihai line, which uses modified spinning, dyeing, and weaving equipment. Gasgoo attributes these figures to the company.
Dealroom additionally reports annual production capacity of 200,000 square meters. Capacity is a potential output rate, however, not a record of material shipped or sold.
These metrics answer different questions:
- Client count describes the reported customer footprint, without revealing order size or repeat purchases.
- Glove shipments indicate delivered hardware, without showing utilization or useful training hours collected.
- Yield describes the share passing a manufacturing acceptance process, whose criteria are not detailed.
Frequently Asked Questions
3 questions
1What Is Fabric Electronic Skin?
It is textile with integrated sensing functions. Juqiao uses a woven fiber network whose electrical response changes under contact.
