NUS Built a Textile Touchscreen You Can Wear and Wash
A fibre that acts as pixel and touch sensor, 5,000 chips and 20,000 connections in a thread. The Nature result is real; the prototype still needs a controller and a battery.
The National University of Singapore has built a textile touchscreen into a sleeve, and the result is not a rigid panel stitched to fabric. It is a thread. The team, led by Tural Khudiyev in the Department of Materials Science and Engineering, developed a fibre that works as both pixel and touch sensor, and wove 54 of them into a sleeve to make a 96 by 54-pixel display. Nature published the work, with a commentary alongside it, and the prototype displays images and video, responds to finger presses, reads body temperature and survives 50 cycles in a washing machine.
The result matters because it attacks a dimensional problem that has held back wearable displays. A screen is a two-dimensional surface. A textile is a collection of one-dimensional fibres. Engineers have made flexible displays and they have made conductive yarns, but nobody had produced a fibre that could be addressed pixel by pixel, sense a finger press along its own length and still behave like thread. That is the claim the paper makes, and the evidence is a working sleeve rather than a rendering.
Difficulty is Geometric
The difficulty is geometric, and the paper is explicit about it. A fibre is a one-dimensional, axially symmetric object, and the tools that make conventional electronics assume a flat, two-dimensional surface. Chips are designed to sit on a board; a fibre has no board. The team’s answer was a new interconnection architecture that turns the fibre’s length into a bus, with a common propagation path running along the axis and each chip tapping into it where needed. The touch layer works differently again: it is coaxial, wrapped around the core, and a press completes a resistive path whose signal tells the controller where along the length the finger landed. Multi-touch came from connecting both ends of the sensing layer and driving them asymmetrically, which is a neat trick rather than an extra component.
What the team built
The fibre is a millimetre thick, and the team draws it ten metres long. Inside it sit addressable LEDs, a sensing layer that runs coaxially along the length, and the wiring that connects them. To make it, the team devised a programmable method for creating connection points along insulated wires exactly where each microchip needs one: a laser exposes the spot, and controlled surface wetting guides molten solder to bond only there.
That let them integrate more than 5,000 chips through over 20,000 electrical connections, then encapsulate the assembly in a multilayer fibre using thermal drawing, a technique borrowed from optical fibre manufacturing that heats a macroscopic preform and pulls it down to a thread while preserving its internal architecture. In the prototype, 54 pixel fibres, each carrying 96 LEDs, form a 168 by 105 millimetre display that refreshes at 30 frames per second with 8-bit colour. That is the textile touchscreen the paper reports.
What the prototype can do
The sleeve runs a calculator, shows a photo gallery, plays video at 30 frames per second, works as a compass, takes voice commands, receives images over Bluetooth and adjusts its own brightness. Alongside the pixel fibres sit peripheral fibres made the same way, carrying a microphone, a speaker and sensors for movement, magnetic north, temperature and ambient light; placed under the arm, the temperature sensor lets the sleeve read and display body temperature. There is a second advantage the release mentions only in passing and the paper makes more of: modularity. Commercial foldable displays are monolithic, so a localised failure means replacing the whole module.
A woven display is a collection of independent fibres, each with its own control channel, so a failed line can be identified, pulled and replaced without rebuilding the panel. For a garment, that is the difference between a repair and a write-off, and it matters when the device is expected to be worn and washed rather than carried in a case.
The system also registers multiple touches at once. The durability claims are the ones that matter for a garment: light output and touch response barely changed after 50 detergent washes and 1,000 bends around a five-millimetre radius, the fabric breathes about as well as cotton or denim, and the materials stay within skin-contact safety limits.
The caveat the announcement leads past
The prototype is not yet a self-contained garment. As independent coverage noted when the paper appeared, the current version relies on separate microcontroller circuits connected to the fibres by cables and on external battery packs. The fibre itself is washable and flexible; the electronics that drive it are not yet part of the weave. That is not a fatal objection, because the paper is a materials and architecture result rather than a product, and the authors say smaller chips and thinner fibres are compatible with the method.
But it is the gap between a laboratory demonstration and a jacket you can wear to work, and it points to the two problems the field has to solve next: getting the controller into the textile, and getting power into it without a battery pack. Independent researchers quoted in the coverage name both, along with flexible energy-harvesting threads that could draw power from body heat or movement.
The wider field
The NUS work sits inside a fast-moving line of research. In February 2025, a group at MIT led by Yoel Fink published a fibre computer that combined sensing, memory, processing and communication in under five grams, was machine-washable and could go into garments; four such fibres, one per limb, classified physical activity at 95 per cent accuracy when networked. Thermal drawing, the technique both teams use, descends from optical fibre manufacturing and can already produce kilometre-long multifunctional fibres, which is why it is the most industrially credible route. In October, researchers at Donghua University published a different approach in Nature Electronics, using a five-channel microfluidic spinner to build circuits inside fibres as they form, and demonstrated gesture control of drones and robotic arms. Reviews of the field put the remaining challenges plainly: encapsulation, interconnect reliability and scalable manufacturing, with practical translation still described as being in its infancy.
The limits
It is worth being precise about what this textile touchscreen is not. At 96 by 54 pixels it is far coarser than any phone screen, and the paper compares its form factor with foldable commercial displays rather than its resolution. It runs video, but not at the frame rates or colour depth a consumer expects from a device. The team tested it for 50 washes, not the hundreds a garment might see. And the 8-bit figure describes intensity control per element rather than the full-colour pixel grid a phone panel produces. None of that diminishes the result, which is a genuine first on the architecture, but a reader should not expect a wearable television.
What to watch
Three things will decide whether the textile touchscreen becomes a product category. The first is integration: can the controller and the power source move into the fibre, so that the garment needs no tether? The second is resolution: the authors say smaller chips and thinner fibres should raise it, and the knitting, weaving and embroidery compatibility of the fibres means standard textile machines could do the assembly. The third is the application that justifies the cost, and the team points at healthcare, where a garment could collect physiological data continuously and let the wearer interact with it directly. That is a more compelling case than a sleeve that plays video, and it is the one that would pay for the electronics.
The healthcare case
The healthcare case deserves more scrutiny than a quotation, because it is where the money would come from and where the obligations are heaviest. A garment that continuously collects physiological data and displays it on the sleeve counts as a medical device in most jurisdictions, which brings requirements for accuracy, validation, data protection and post-market surveillance that a research prototype does not yet address. It also raises the question the field keeps deferring: what happens to the data, who can access it, and what the wearer is told. Wearable health devices already face this: a fitness tracker that estimates heart rate carries lighter obligations than a garment a clinician could rely on, and the line between the two is a regulatory judgement rather than a technical one. The NUS team has demonstrated the sensing and the display. The regulatory pathway is a separate body of work.
What it signals
The broader signal is that the boundary between electronics and textiles is dissolving, and it is dissolving from the fibre up rather than from the screen down. The textile touchscreen is the clearest evidence of it. For a decade, wearable computing meant a rigid device strapped to a wrist. The direction of this work is different: the fabric itself becomes the interface, the sensor and the bus, and the electronics hide inside the thread. That has obvious appeal for health monitoring, for sportswear and for anyone who has wanted to stop reaching for a phone.
It also raises questions the field has not answered: what happens to a medical device that goes through the wash, who repairs a fibre when it fails, and what data a garment collects and where it goes. The NUS team has shown that the physics and the manufacturing can work. The governance and the product design are still ahead.