NVIDIA and Zalando Integrate Cosmos for 48-Hour 3D Product Pipeline

At a glance

Zalando has deployed an integrated AI pipeline powered by NVIDIA Cosmos, Gen-3C and DiffusionRenderer to convert physical products into high-fidelity 3D assets ready for digital storefronts within 48 hours.

What changed

The company integrated ALLSIDES’ photometric 3D scanning platform with NVIDIA’s AI models. The resulting workflow automates capture, reconstruction and rendering at scale, replacing slower manual or semi-automated 3D content creation processes.

Why it matters

Operationally the new pipeline reduces production time from weeks to 48 hours, lowering cost and accelerating catalog refresh cycles. Commercially it enables richer product visualization that supports higher customer confidence and conversion rates in fashion e-commerce. From a compliance perspective the governed NVIDIA-based pipeline improves consistency and auditability of generated assets used across regulated digital commerce channels.

Key details

• Photometric scanning captures accurate geometry and material properties.
• NVIDIA Cosmos, Gen-3C and DiffusionRenderer generate production-ready 3D models.
• End-to-end cycle from physical scan to live storefront now completes in two days.
• Deployment targets high-volume product catalog requirements at Zalando.

Sources

• https://x.com/nvidia/status/2057872121984930177
• https://x.com/NVIDIAAI/status/2057855521193881773

Notes for citation

All quantitative claims (48-hour cycle, platform names, model components) are taken directly from the NVIDIA announcement. No external benchmarks or performance numbers were added.

Want to discuss how this affects your workflows? Book a call →

AI-assisted analysis by Skirr AI