Nvidia and Nokia to Pioneer the AI Platform for 6G (nvidianews.nvidia.com)

Nvidia and Nokia announced a collaboration to build an AI platform tailored for 6G networks that combines Nvidia’s GPU-accelerated AI stack — including the DoMINO NIM microservice and the PhysicsNeMo AI physics framework — with Nokia’s telecom systems and radio access expertise. The partnership aims to move beyond conventional ML workflows by embedding physics-aware foundation models and GPU-accelerated simulation/learning into the cloud–edge continuum, enabling use cases such as real‑time RAN optimization, network digital twins, and faster engineering design cycles already being piloted in aerospace and automotive. For the AI/ML community this is significant because it accelerates the integration of physics-informed models and large-scale, low‑latency inference into operational networks. Key technical elements include microservice-based model deployment (DoMINO NIM) for scalable distributed inference, GPU-accelerated physics simulation and neural operators (PhysicsNeMo) for tighter simulation-to-ML loops, and targeted edge/cloud co‑deployment to meet 6G latency, reliability, and energy constraints. The collaboration will likely drive new benchmarks, MLOps patterns for distributed telecom AI, and research into model compression, partitioning and hardware–software co‑design needed to operationalize physics‑aware foundation models at telecom scale.