Fashion and spatial computing have had an awkward relationship. The technology has been hyped repeatedly — virtual fitting rooms were a CES fixture for years before anything meaningful shipped. But the last two years have produced real deployments with real results, driven by improvements in 3D garment simulation, on-device AI, and the emergence of platforms that make virtual try-on practical to implement at scale.
This is where the industry actually is in mid-2026.
What Virtual Try-On Actually Looks Like Now
Consumer-facing virtual try-on splits into two distinct categories that have very different maturity levels.
Accessories and cosmetics AR — trying on sunglasses, jewellery, watches, lipstick, eyeshadow — works well and has for several years. The reason is geometry: accessories sit on the surface of the face or wrist without interacting with the body, so they can be anchored to AR landmarks (eye corners, wrist bones detected by phone cameras) with convincing results. Snap AR, Perfect Corp, and WANNA have deployed these experiences at scale with major brands including Gucci, Dior, and Rolex.
Clothing try-on is harder by an order of magnitude. Fabric drapes, stretches, and deforms based on the wearer’s body shape and movement. Accurately simulating how a garment fits a specific person requires knowing their body dimensions, the fabric properties, and the garment’s construction — then running a physics simulation in real time. Early attempts either looked obviously artificial (the clothing didn’t move with the body) or required the user to upload body measurements and wait for a server-side render.
The state of the art in 2026 is better but still limited. Platforms like WANNA (focused on knitwear and outerwear), Zeekit (now integrated into Walmart), and CLO Virtual Fashion’s consumer-facing tools can produce plausible simulations for relatively simple garments — T-shirts, knitwear, denim. Complex tailoring, formal wear with structured construction, or garments with unusual cut or fabric drape still produce results that look visibly wrong to a discerning eye.
The business case even for imperfect clothing try-on is real: studies from Shopify and various retailers consistently show reduced returns when shoppers use try-on features before purchasing, and returns in apparel (15–30% of online purchases) are a significant cost.
Digital Showrooms for B2B Fashion
The B2B use case — brand to retail buyer — has seen faster and more complete adoption than consumer-facing try-on.
Fashion brands traditionally show upcoming collections to retail buyers through physical showrooms and seasonal trade shows. This requires producing physical samples (expensive), transporting them, and scheduling buyers to attend physical locations. The pandemic accelerated digital showroom adoption, but early solutions were essentially 360-degree photo carousels — not meaningfully immersive.
Today’s digital showrooms use 3D models of garments (often created by the same CLO or Browzwear software used for digital product development) displayed in browser-based or native XR viewers. Buyers can:
- View garments in all colourways without physical samples existing
- Mix and match separates to see how they work together
- See garments modelled on different body types
- Place orders directly within the platform
Brands including Tommy Hilfiger, Hugo Boss, and Stella McCartney have published adoption of digital showrooms. The efficiency case is significant: reducing physical sample production by even 30–50% is meaningful in an industry where a single sample can cost hundreds to thousands of pounds, and a collection might require hundreds of samples.
The VR showroom — walking through a digital space and interacting with hanging garments using a headset — remains a niche within this. Most digital showroom deployments are browser-based or tablet-based, not VR-headset-based. The friction of requiring buyers to wear a headset for a business meeting remains real.
Luxury and Brand Experience
Luxury fashion brands have used immersive experiences as marketing and brand-building for longer than mass-market retailers. This is where the more experimental spatial computing deployments have happened.
Burberry, Gucci, and Louis Vuitton have all deployed AR experiences in various forms — branded AR filters, in-store AR try-on with dedicated hardware, and standalone immersive brand experiences. These tend to be high-production-quality one-off projects with a limited release window rather than persistent product features.
The most interesting deployments in 2026 are in physical retail: in-store AR mirrors that let customers see how an outfit looks in different colours or styles without physically changing, or interactive window displays that respond to passers-by. These typically run on dedicated hardware rather than customer smartphones, which allows for higher production quality and removes the “download the app” friction.
The Apple Vision Pro has been used for a small number of luxury brand experiences — notably viewing jewellery or accessories in high-fidelity 3D in a spatial environment. The use case maps well to Vision Pro’s strengths (high-resolution display, precise hand tracking) and the target demographic (customers for whom a £3,000 headset is plausible). Whether this remains a marketing novelty or becomes an ongoing sales channel depends on Vision Pro’s penetration into the luxury consumer segment.
3D Garment Creation Tools
Underlying all of the above is 3D garment creation software that lets designers work digitally from the start rather than digitising physical samples after the fact.
CLO 3D and Browzwear are the two dominant tools for pattern-based garment simulation. Both simulate fabric behaviour with physics engines that account for drape, stretch, weight, and construction details. Designers can create and iterate on digital prototypes before any physical sample is cut, reducing development cycles.
Optitex targets the mass-market end of the market with deeper integration into existing product lifecycle management (PLM) systems. Adobe Substance (following the Fabric Engine acquisition) provides procedural material tools for generating photorealistic fabric renders.
The workflow that’s emerging in larger fashion brands is: design digitally in CLO → render photorealistic images for catalogue — export to GLTF/USD for use in digital showrooms and consumer-facing AR → produce only the final production samples for quality checking. This reduces total sample count while maintaining the ability to review and iterate.
Where the Challenges Are
Fit accuracy: size standardisation is poor across the industry. Even with accurate body measurements, how a garment “fits” depends on construction details that are difficult to encode in a simulation.
Returns from virtual try-on: early data is mixed. Some deployments show reduced returns; others show no meaningful difference, possibly because consumers try more unusual purchases with the safety net of virtual try-on before buying. Measuring incrementality is genuinely difficult.
Consumer adoption friction: downloading an app, enabling camera access, and positioning yourself in frame is still friction, even if less than it was. WebAR (AR running in a browser without an app install) has improved substantially and is increasingly used for fashion try-on.
Sustainability accounting: digital showrooms and reduced physical samples represent genuine sustainability improvements. The industry’s sustainability credentials are heavily scrutinised, and the digital product development argument — fewer samples produced, shipped, and discarded — is real and increasingly used in brand communications.
The Direction of Travel
Virtual try-on for accessories works and is deployed at scale. Virtual try-on for clothing works for simple garments and is improving. Digital showrooms for B2B buying are genuinely adopted and delivering ROI for major brands.
The next step change will likely come from AI-driven body shape estimation (making try-on require less setup from the consumer) and from improvements in real-time fabric simulation on device. Both are active research and engineering areas, and the trajectory from 2024 to 2026 suggests meaningful progress rather than stagnation.
Fashion is not the easiest domain for spatial computing — the core problem (how does this garment look on this specific body?) involves hard physics and hard computer vision. But the commercial incentives are large enough that the tooling will continue to improve.