AI-RAN technology is transitioning from laboratory trials to live field deployments, signalling a shift toward software-defined, AI-native wireless networks. Ahead of Mobile World Congress (MWC) in Barcelona (March 2–5), NVIDIA and Nokia announced expanded AI-RAN collaborations with leading telecom operators across Europe, Asia and North America, powered by NVIDIA’s AI-RAN platforms.

Major operators, including T-Mobile US, SoftBank and Indosat Ooredoo Hutchison, have achieved key milestones, taking NVIDIA-powered AI-RAN outdoors and over the air. New benchmarking results from SynaXG demonstrated carrier-grade reliability across multiple 5G spectrum bands using NVIDIA platforms, achieving 36 Gbps throughput and under 10 milliseconds latency on a single NVIDIA GH200 server, including the world’s first AI-RAN implementation on millimetre wave (FR2) bands.

In live field tests, T-Mobile demonstrated concurrent AI and RAN processing using Nokia’s CUDA-accelerated RAN software and AirScale massive MIMO radios. SoftBank achieved an industry-first 16-layer massive MIMO using fully software-defined 5G on NVIDIA’s AI-RAN platform. Meanwhile, Indosat showcased Southeast Asia’s first AI-powered 5G call, enabling real-time robotic control over a live 5G network.

At MWC, over 20 AI-RAN Alliance demos built on NVIDIA AI Aerial platforms will highlight innovations ranging from AI-native air interfaces by DeepSig to GPU-sharing orchestration blueprints from zTouch Networks and AI switching solutions developed by Northeastern University and SoftBank.

Industry leaders say the convergence around GPU-accelerated, software-defined architectures is laying the groundwork for secure, open and AI-native 6G systems. As AI workloads expand across edge and cloud environments, AI-RAN is emerging as a unifying architecture designed to support next-generation autonomous systems, robotics and intelligent mobility at scale.