NVIDIA and Cisco Enable Edge AI for Retail
Cisco and NVIDIA are partnering to accelerate the deployment of edge AI in retail for applications like real-time video analytics and shelf monitoring. The hybrid approach balances on-premise processing on NVIDIA-powered devices with cloud-based orchestration. Separately, Grid Dynamics unveiled an NVIDIA Solution Center to provide modular, production-grade AI solutions for retail and manufacturing.
- The global retail edge computing market was valued at USD 2.15 billion in 2024 and is projected to reach USD 4.16 billion by 2032, growing at a CAGR of 8.60%, driven by the demand for real-time data processing from IoT devices and AI analytics. - The underlying NVIDIA technology for many of these in-store video analytics applications is the NVIDIA Metropolis platform, which provides a framework and developer tools for vision AI. For example, the AI workflow for retail loss prevention uses Metropolis microservices and models pretrained on the most frequently stolen goods to provide intelligent alerts. - The Grid Dynamics NVIDIA Solution Center leverages specific NVIDIA software like Metropolis for video analytics, Omniverse for digital twin simulations of warehouses, and Nemotron for content generation to create deployable solutions. These were jointly presented by the companies at the National Retail Federation (NRF) conference in January 2026. - This retail-focused effort is part of a broader Cisco and NVIDIA partnership to deliver simplified AI infrastructure. The collaboration includes integrating Cisco's Silicon One networking technology into NVIDIA's Spectrum-X Ethernet platform to create unified architectures for AI data centers. - A key driver for these solutions is loss prevention, a problem that costs retailers around $100 billion annually. According to one report, advanced AI-powered loss prevention systems can deliver a 374% return on investment over three years. - The hybrid edge-cloud architecture is critical for applications requiring immediate local processing to reduce latency. By processing data within the store, retailers can enable real-time applications like point-of-sale systems and instant promotions without relying on constant internet connectivity to a central cloud. - Use cases extend beyond asset tracking to include creating digital twins of warehouses using NVIDIA Omniverse. This allows logistics and manufacturing clients to simulate and test optimal facility layouts and inventory picking paths without costly physical reorganizations. - One of the stated goals for the Grid Dynamics NVIDIA Solution Center is to provide an alternative to recurring SaaS licensing fees. The center offers enterprises high-performance AI applications that they can deploy on-premise, at the edge, or in the cloud, giving them more control over costs and infrastructure.
