CATCHES Opens to Retailers with PredictiveFit Virtual Try-On, Launching on THE OUTNET, After RealFit Lifted AMIRI Online Conversion by 55%

• CATCHES launches PredictiveFit at NVIDIA GTC Berlin, bringing physics-based AI try-on to retailers for the first time
• THE OUTNET is the first retailer to go live with CATCHES, from October 21, 2026
• PredictiveFit technology builds photorealistic try-on visuals from the product data retailers already hold
• Six months after AMIRI pioneered CATCHES’ RealFit technology at NVIDIA GTC San Jose, activated styles saw units sold up 154% and gross sales up 156%
• After adjusting for seasonality against comparable styles, conversion rose 55%, add-to-cart 80% and gross sales 56%
• CATCHES is backed by investors including Antoine Arnault and Natalia Vodianova

BERLIN--(BUSINESS WIRE)--CATCHES, the deterministic AI company building the application and infrastructure layers for digital fashion, today launched PredictiveFit, opening its sizing and virtual try-on technology to retailers for the first time. THE OUTNET is the first retailer to go live, from October 21, 2026. CATCHES also released the first performance results from its RealFit launch with AMIRI, six months after the technology debuted at NVIDIA GTC San Jose in March 2026.

RealFit, launched in March, is built for brands. It uses data only the brand holds, such as patterns, fabric specifications and grading, to simulate each garment with millimeter-level fidelity. Retailers sell products they did not make, so they do not have that data. PredictiveFit closes the gap. It builds from product data a retailer already holds, and gives shoppers the same experience: a photorealistic digital twin of their own body and a view of how each garment fits, drapes and moves in each size before they buy.

Today’s launch on THE OUTNET marks CATCHES opening for business with retailers. The technology is live now and can be seen on a THE OUTNET style at catches.ai/outnet-live-style.

“Every retailer is being offered VTO, but CATCHES takes the customer experience to another level. Its unique physics-based, deterministic AI models the shopper, the garment and how the fabric behaves, so customers can see how clothes will look and fit on their own bodies.

“For 25 years, buying fashion online has often meant best guessing your size depending on the designer ordering two and sending one back. CATCHES fundamentally changes that experience. Customers can choose with confidence and enjoy shopping without all the unnecessary back and forth reducing customer friction and unnecessary touch points. They find something they love, it arrives, it fits, and it’s what they expected. That’s what getting the customer journey right looks like.

“For us, that means happier customers, stronger engagement, higher sales and fewer returns. The return on investment follows from CATCHES delivering a better experience. We’re proud to be the first retailer to bring this pioneering technology to our customers.”
- Joe Edery, Founder & CEO, Icon Luxury Group (Co-owner of THE OUTNET)

RealFit results with AMIRI

RealFit lets shoppers enter their measurements and a photograph to create a digital twin, then see with mirror-like realism how each garment fits, drapes and moves on their own body, toggling between sizes before they buy.

Comparing the six months before activation with AMIRI to the six months after, the RealFit styles delivered:

• Conversion rate up 69% (55% seasonally adjusted)
• Add-to-cart up 96% (80% seasonally adjusted)
• Units sold up 154% (39% seasonally adjusted)
• Gross sales up 156% (56% seasonally adjusted)
• Fit-related returns down 30%

The comparison measures each RealFit style against the performance of comparable AMIRI styles without RealFit over the same period, isolating the effect of the technology from seasonal demand. On that conservative basis, shoppers using RealFit were more than half as likely again to buy, and add-to-cart rates almost doubled.

“When we launched at GTC we said fit uncertainty was the biggest problem in fashion e-commerce. Six months later the data from AMIRI shows what happens when you solve it. Until now that was only possible for brands with their own data. PredictiveFit changes that, and THE OUTNET is the first retailer to bring it to its customers. What’s fascinating about these numbers is that we’ve still got a lot of UI to optimize, so they are conservative. We’re really excited about where this can go and what it means for the whole industry,” said Ed Voyce, Founder and CEO, CATCHES.

“The results CATCHES has delivered with AMIRI are exactly what physically grounded technology looks like in practice - not a generic recommendation, but a simulation of how a real garment moves on a real body, built on NVIDIA Omniverse, Cosmos and Newton - converting that fidelity directly into a 55 percent lift in conversion,” said Azita Martin, Vice President and General Manager of AI for Retail and CPG, NVIDIA. “PredictiveFit now brings that same capability to any retailer, at catalog scale. This is the moment physics-based simulation moves from a feature retailers test to infrastructure they run their business on.”

CATCHES now offers two routes onto the platform. Brands use RealFit, which delivers millimeter-level fit fidelity for garments digitized with the brand’s own manufacturing data. Retailers use PredictiveFit, built on the CATCHES Automatic Garment Pipeline, which turns retail product data into simulation-ready OpenUSD garments at catalog scale. In both cases, CATCHES’ sizing and styling model turns every result into a size recommendation and a plain-language fit explanation for the shopper.

For brands and retailers, CATCHES offers a measurable route to higher conversion and fewer fit-related returns, the leading driver of reverse logistics costs in fashion e-commerce. Brands and retailers interested in RealFit or PredictiveFit can reach CATCHES via the “Get in touch” button at catches.ai.

The experience is built on NVIDIA CUDA and Omniverse libraries, leverages the Newton Physics Engine, and combines custom-trained diffusion models with fine-tuned VLMs, including AI models based on NVIDIA Cosmos world foundation models. By building the solution on an OpenUSD foundation, data can seamlessly exchange between the digital twin model, Blender, and Unreal, while NVIDIA RTX Ray Tracing, DLSS, and DLAA deliver real-time, high-fidelity output. CATCHES’ bare-metal infrastructure is distributed across NVIDIA RTX PRO™ 6000 Blackwell Server Edition, NVIDIA Hopper and NVIDIA Blackwell and Blackwell Ultra GPUs, provided at high scale through infrastructure partners Nebius, Scan Computers and Zero Latency (0.lat).

About CATCHES

CATCHES pioneers physics-backed deterministic AI and agentic ecommerce for personalized garment fit across fashion e-commerce, transforming how premium brands deliver digital fashion experiences. The founding team includes PhDs who previously worked on simulation technology at Disney and Pixar. CATCHES is backed by investors including Antoine Arnault and Natalia Vodianova. CATCHES is an NVIDIA Independent Software Vendor for the fashion industry and is based in New York and London. Website: catches.ai

About AMIRI

Established in Hollywood in 2014, AMIRI has led a unique path in luxury fashion, steeped in Americana and shaped by the vision of its namesake Founder and Creative Director, Mike Amiri. AMIRI began by creating exceptional, one-off stage pieces for legendary Hollywood performers and today continues to celebrate a new brand of American luxury, a meeting point of music, art, and fashion. Now a global brand, yet with a distinct Californian identity, AMIRI is one of the fastest growing businesses in luxury fashion. Website: AMIRI.com

About THE OUTNET

Founded in 2009, THE OUTNET is a leading online multiband luxury fashion retailer carrying more than 350 designers. Its curated selection of womenswear and menswear includes exclusive capsule collections. Serving customers in more than 100 countries, THE OUTNET makes designer fashion more accessible while extending the life of luxury beyond the current season. THE OUTNET was acquired by Icon Luxury Group in the first quarter of 2026. Website: theoutnet.com