Updated: 4 days ago
History, Present, and Possibilities of the Digital Twin in Fashion: From Apollo to Zara's Fitting Room, to the New Technology Launched by Catches
Every fitting room in the world holds the same question: How does it look on me? Not "What is this dress like?" The photograph already tells you that. It's about how it falls on my body, how it moves when I walk, whether it pulls at the shoulders, and whether the waist sits where it should. This is the question e-commerce has never been able to answer. Today, it produces one of fashion's quietest costs: returns. In some categories, they exceed half of all purchases, almost always for the same reason—the wrong size.
The answer the industry is betting on has a name borrowed from aerospace engineering: the digital twin.
A Story That Begins in Space
The idea of a double was born in the 1960s with the Apollo programme. NASA built vehicles identical to those in flight and kept them on the ground. When something went wrong in space, engineers recreated the conditions on the physical twin to work out how to help the crew. In 1991, computer scientist David Gelernter imagined, in his book Mirror Worlds, digital worlds that reflect real ones.
The concept proper arrived in 2002. Michael Grieves, a scholar of product lifecycle management, presented it at an engineering conference in Troy, Michigan. He described a physical object, its virtual copy, and a channel that keeps them in dialogue. The name "digital twin" was coined in 2010 by John Vickers, a NASA researcher, in a document on the technologies of the future.
For fifteen years, the digital twin lived among jet engines, factories, and satellites. Then it met the body.
The Long Dress Rehearsal
Virtual fitting rooms have existed for more than a decade. Start-ups such as Fits.me, Metail, and Virtusize tried to solve online sizing from the early 2010s. They used robotic mannequins, avatars, and comparison charts. In 2013, Israeli founder Yael Vizel launched Zeekit, which mapped a person's photo into thousands of segments to "dress" it digitally. Brands like Asos, Adidas, Macy's, Tommy Hilfiger, and Levi's used it. In May 2021, Walmart bought it and integrated it into its app in 2022. The company claimed a 36% reduction in returns.
Today, the gesture has gone mainstream. Since mid-December 2025, Zara has offered virtual try-on in its app. Users upload a photo of their face and a full-length one. The app generates an avatar wearing the chosen piece, shown in motion too. Within a few months, it was live in 43 markets with more than seven million sessions. Zalando and Revolve have launched similar features, and Google offers virtual try-on directly in search. When the world's largest fast-fashion brand adopts a technology, it stops being a curiosity and becomes an expectation.
Beautiful Is Not Enough
There is, however, a limit that anyone who works with images knows well. Most of these tools are image generation. Artificial intelligence produces a plausible photo of you wearing the garment. It is convincing, often beautiful, but it does not really know your measurements or those of the dress. It shows how it might look, not how it would fit.
This is where Catches comes in, a company based between London and New York, founded by Ed Voyce. The founding team includes PhDs who previously worked on simulation technology at Disney and Pixar. Its starting point is almost provocative: according to Voyce, many of these models are lovely to look at but help no one choose a size. Before applying any artificial intelligence, Catches spent two years building a physics simulation engine, studying real fabrics. The result is a digital twin in the strict sense: a body that matches yours and a garment that behaves like that fabric, with its weight, structure, and the way it falls and moves.
The first version, RealFit, debuted in March 2026 at Nvidia's GTC conference in San Jose, on Amiri's website. It uses data only the brand holds: patterns, fabric specifications, and size grading. After six months, the numbers speak for themselves. Compared with similar styles without the technology, and adjusted for seasonality, conversions grew by 55%, add-to-cart by 80%, and gross sales by 56%, while fit-related returns fell by 30%.
Behind it is a $10 million round with names that say a great deal about the stakes: Antoine Arnault, Natalia Vodianova, former Tommy Hilfiger CEO Gary Sheinbaum, and Sarah Willersdorf, BCG's former head of luxury.
This Week's News
On 22 October, at GTC in Berlin, Catches took the next step. It is called PredictiveFit and is designed for those who sell many brands and do not own each one's patterns. An automatic system turns the product data a retailer already holds—meaning product sheets, measurements, and photographs—into digital garments ready for simulation at catalogue scale. The first to use it is The Outnet, the luxury outlet acquired in 2026 by Icon Luxury Group.
The same day saw the birth of Casi, a personal assistant that lives inside WhatsApp. It searches, pairs, and finds the right size through conversation, like a sales assistant who knows your measurements.
What Becomes Possible
For shoppers, the digital twin shifts the question from "Do I like it?" to "Does it fit me?" You can see the same jacket in two sizes, understand how a long dress moves, and decide without ordering three sizes to send two back. For brands, it means fewer returns, fewer pointless shipments, and therefore less waste. It also fosters a more intimate relationship with the customer because the garment is tried on their body, not on a model's.
For those who build images, it is a paradigm shift. The digital garment no longer only has to appear; it has to behave. Fabric becomes data, drape becomes calculation, and precision in rendering the material stops being an aesthetic detail and becomes the condition of the sale.
There is a more distant prospect, one Grieves intuited twenty years ago: the digital twin can precede the real object. A garment could be tried on, corrected, and even sold before it is sewn. Fashion, which has always lived a season ahead, may find here its most natural tool.
For centuries, the mirror has given us back an image. The next one, perhaps, will also know what the dress is made of.
Sources: Catches press releases (Business Wire, 16 March and 22 October 2026); WWD, 21 October 2026; Nvidia, 22 October 2026; FashionNetwork, March 2026; CNBC, April 2026; Inditex FY2025 results; Walmart, 2022; on the history of the digital twin: VentureBeat, Advanced Manufacturing, NASA.
Vertigini Journal is published by Vertigini Studio, a Milan studio creating AI imagery for fashion, beauty, and jewellery brands. Discover our E-commerce & Lookbook work.
Comments