In the basement of an ordinary house, during lockdown, a professor tinkered with 3D-printed parts and brass tubes. Now he's doing it in a lab. He's not assembling a drone, but a grid of needles whispering to the thread: move forward, move sideways, add volume. The 3D fabrics take shape, layer by layer, like a pyramid growing from nothing.
François Guimbretière, of Cornell University, observes: her prototype doesn't weave sheets, but solid objects with calculated rigidity. It makes almost intimate gestures, this loom that becomes a sculptor. But will it withstand the chaos of the thread? Curiosity remains there, hanging on a crochet hook.
But then the thread stumbles
It happens. The loom head slips on the 6x6 grid, catches an extra stitch, or loses one. The prototype, developed between Cornell and Carnegie Mellon, is still in its infancy: it creates a curved C, a crochet pyramid, but sometimes the thread tangles or falls off. Guimbretière knows this: "We work slowly, and every incorrect loop is a reminder." The system adds stitches in every direction (forward, backward, diagonal), but the speed is that of a craftsman, not a factory.
Yet this is precisely where the potential for 3D fabrics lies: complete control over the structure, without using molds or glue.
How the trick works
At the heart of the project is a cluster of symmetrical needles, each with a 3D-printed double hook and supported by a brass tube. A custom code generates the digital pattern, and the head moves freely across the grid and builds layer upon layer, like a 3D printer but with a wool core. The result? 3D fabrics with soft and hard zones on command, seamlessly. The team presented their work at UIST 2025 in Busan (UIST 2025), showing that solid knitting is doable. Simply increase the needle count to scale, says Guimbretière.
A technical detail? The code is open: anyone can generate patterns for complex shapes. A bit like a CAD for sweaters, but which prints in volume (Near Future, 2024).
3D fabrics, the thread paradox
I thought about it, and I found something that didn't quite fit. On the one hand, this system promises textile prosthetics, artificial blood vessels, and structures that grow with cells. On the other, the prototype is still slow, fragile, and loses stitches like a beginner.
It's a bit like a surgeon with his hands tied: he knows where to cut, but the needle slips. 3D medical tissues are the goal, but they require absolute precision, and the thread doesn't always cooperate. Guimbretière admits: "We need to make the system robust before considering a ligament." The technology is there, but the humanity of the thread (that random knot) reminds us that not everything can be planned. We need more time, come on, let's not pretend.
3D fabrics, a fabric that waits
3D fabrics are there, tangible but not yet ready. Adding needles is easy, removing knots is less so. Guimbretière talks about medical applications, but he knows the road is long. As with any good prototype, the promise surpasses reality: woven veins, bone supports, materials that adapt to the body. A recent study on Textile Research Journal confirms that programmable structures of this type are the future (Textile Research Journal, 2024). Solid knitting exists. Now it just needs to learn not to trip over its own thread.
