Meet Netris at GTC 2026

AI workloads are scaling into full production for training and inference. Standing up these clusters for R&D is challenging on its own, but production brings complexity: stricter security requirements, operational rigor, and a broader user base demanding stability, flexibility, and mature features. The solution is a single, enterprise-grade, software-defined NVL GPU cluster built to dynamically adapt to changing workloads, support rapid deployment, and simplify life-cycle management. In this talk, we’ll examine considerations across the infrastructure, platform, and software layers. We’ll look at network automation, virtualization, and different orchestration platforms. We’ll focus with a concrete example based on our reference architecture, integrating Run:ai, Base Command Manager, vCluster, and Netris to create a secure, scalable, production-ready multi-tenant cluster.