NVIDIA Powers the Future of Fashion E-Commerce with CATCHES' Generative AI Launch

In a groundbreaking press release on March 16, 2026, CATCHES, a pioneering physics-backed AI company, introduced its innovative Generative AI with Sizing Technology, branded as "RealFit." The launch took place at NVIDIA's GTC conference, showcasing a transformative new approach to online fashion shopping that promises to redefine customer experiences in e-commerce.

RealFit allows online shoppers to visualize how garments will fit, drape, and move on their bodies with astonishing accuracy, akin to a mirror. By entering personal body measurements and uploading a photograph, users can create a digital twin and see how clothing from the AMIRI collection fits them. The technology not only allows customers to toggle between different sizes but also addresses a persistent challenge in the fashion industry—fit uncertainty. This issue has historically led to high return rates, with over 50% in certain categories, thus impacting online conversion rates.

The first public deployment of the RealFit technology is now live on the AMIRI website. Mike Amiri, the founder and creative director of AMIRI, emphasized the synergy between CATCHES’ innovation and the brand's core values of community, craft, and innovation. He highlighted how the RealFit technology not only personalizes the shopping experience but also strengthens the connection between AMIRI and its customers, whether they are shopping online or in-store.

NVIDIA's involvement in this venture is pivotal. Azita Martin, VP and GM of AI for Retail and CPG at NVIDIA, pointed out that accelerated computing is unlocking new possibilities in digital fashion experiences. CATCHES' RealFit, built on the NVIDIA Omniverse and CUDA platform, leverages advanced physics-based simulations for fabric movements and sizing. This collaboration is set to extend these capabilities to more fashion brands in the coming months.

Ed Voyce, CEO and Founder of CATCHES, addressed the significant challenge of customer fit uncertainty in fashion e-commerce. He stated that the RealFit technology provides a personalized, photorealistic experience that accurately reflects real materials and sizing. This innovation not only aims to enhance customer satisfaction but also offers brands a customized solution that can lead to higher purchase conversions and lower return rates—two critical factors in the e-commerce landscape.

CATCHES has garnered significant backing from high-profile figures in the luxury fashion sector, raising $10 million from investors including Antoine Arnault of LVMH and supermodel Natalia Vodianova. Arnault emphasized the importance of understanding the relationship between garments and their wearers, while Vodianova praised CATCHES for anchoring AI to the laws of physics, marking a significant evolution in digital fashion technology.

The development of RealFit involved two years of research to create a proprietary GPU-accelerated simulation framework. This framework is designed to accurately represent fabric weight, structure, drape, and movement. The technology utilizes a combination of NVIDIA's advanced tools, including CUDA and Omniverse libraries, the Newton Physics Engine, and AI models specifically tailored for fashion applications.

CATCHES' infrastructure is built upon NVIDIA's RTX PRO™ 6000 Blackwell Server Edition and incorporates various GPU models to deliver high-scale performance. The use of open-source USD standards facilitates seamless data exchange between digital twin models and other creative tools, enhancing the overall user experience.

With the launch of CATCHES' RealFit technology, powered by NVIDIA, the future of fashion e-commerce looks promising. By merging cutting-edge AI with physics-based simulations, CATCHES is set to revolutionize how consumers shop for clothing online. This innovation not only addresses long-standing industry challenges but also fosters deeper connections between brands and their customers, paving the way for a new era in digital fashion experiences.