Nokia is running AI-RAN trials with BT, Deutsche Telekom, Elisa and Orange, while nine operators across Europe, the Middle East and Africa are building or expanding NVIDIA-powered infrastructure. Nearly 20 operators have publicly announced AI-RAN collaborations built on Nokia's anyRAN software and NVIDIA's Aerial RAN Computer.
The reference design combines Nokia's anyRAN software with the Aerial RAN Computer, CUDA, Nemotron open models and an Agent Toolkit. Those components are being embedded in Nokia's Autonomous Networks Suite and in third-party operations software from Cisco, TCS and NTT DATA. Dell, QCT and Supermicro supply ARC-ready servers, with Red Hat OpenShift providing the cloud-native foundation.
Trials test a mix of AI workloads
Elisa completed Europe's first AI-RAN field trial, running voice, mobile data and vision AI together on one accelerated platform over 5G. Orange is introducing voice intelligence services for enterprises and emergency responders. Deutsche Telekom is working on physical AI for industrial customers, and BT is preparing live trials in the UK. The work is a workload mix rather than a single radio feature.
The engineering case rests on assets operators already control: land, power, building shells, spectrum and fibre. Adding accelerated compute at centralised, metro and cell sites can place models nearer to machinery, responders and handsets, while allowing the radio itself to run inference. Lab measurements from Deutsche Telekom and Nokia put the typical uplink throughput improvement from AI-based channel estimation at around 10 to 15 percent.
Economics depend on backhaul and revenue
That is a modest gain for a distributed compute build-out. Backhaul becomes the binding constraint once inference moves outward, and Nokia's 800G coherent pluggables indicate the inter-site capacity the company expects operators to need. Until AI-service revenue materialises, operators carry the financing for compute and backhaul, while suppliers of accelerated compute and RAN software occupy the layer where incremental value would accumulate.
The stack shortens integration by combining anyRAN, ARC, CUDA and Nemotron in one reference design, but it also ties operators to NVIDIA's programming model at the radio layer, with Nokia's RAN software above it. Cisco and Nokia are both folding Nemotron open models into autonomous-network software, extending that dependency into operations tooling.
NVIDIA founder and CEO Jensen Huang said telecom networks built to carry conversations would also carry intelligence, and that operators are turning land, power and connectivity into distributed AI infrastructure. Nokia Chief Technology and AI Officer Pallavi Mahajan said the goal is a software-defined foundation that helps operators get more from spectrum today while opening a path toward AI-native 6G.
Disclosed checkpoints cluster in 2027. BT's Birmingham live trial is slated for then, alongside Deutsche Telekom's field trials and Elisa's distributed deployment with commercial AI-RAN services.