Nokia is running trials of its anyRAN software with BT, Deutsche Telekom, Elisa and Orange, as the vendor and NVIDIA push AI-RAN out of reference designs and into live 5G networks. Nearly 20 operators have publicly announced AI-RAN collaborations built on Nokia's anyRAN software and NVIDIA's Aerial RAN Computer, and nine operators across Europe, the Middle East and Africa are now building or expanding NVIDIA-powered infrastructure.
One reference design, several suppliers
The stack pairs Nokia's anyRAN software with the Aerial RAN Computer, with CUDA, NVIDIA's Nemotron open models and its Agent Toolkit embedded both 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, and Red Hat OpenShift provides the cloud-native foundation underneath. Nokia's case is that assembling these parts into a single reference design shortens the integration path for operators.
Where the compute sits
The engineering rationale rests on assets operators already control: land, power and shell space at sites, plus spectrum and fibre. Adding accelerated compute at centralised, metro and cell sites lets operators serve models closer to the machinery, responders and handsets that consume them, and allows 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 roughly 10 to 15 percent.
On the figures disclosed, that is a modest return for a distributed compute build-out. Backhaul becomes the binding constraint once inference moves outward from the core, and Nokia's 800G coherent pluggables hint at the inter-site capacity the company expects operators to need.
What the four operators are testing
- 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.
- BT is preparing live trials in the UK, with a Birmingham trial slated for 2027.
Across the announced projects the work spans voice, mobile data and vision AI rather than a single radio feature, and the trials are being framed as a mix of workloads sharing one accelerated platform.
Jensen Huang, founder and CEO of NVIDIA, described the shift as networks moving from carrying conversations to carrying intelligence, casting operators' land, power and connectivity as a distributed AI infrastructure that brings compute nearer to the industries, people and robots using it. Pallavi Mahajan, Nokia's chief technology and AI officer, said the goal is a trusted, software-defined foundation that helps operators extract more from spectrum today while opening a practical path toward autonomous, AI-native 6G.
Dependencies and who pays
The arrangement also ties operators to NVIDIA's programming model at the radio layer, with Nokia's RAN software sitting above it. Because Cisco and Nokia are both folding Nemotron open models into the software used to run autonomous networks, that dependency extends into operations tooling as well. Until AI-service revenue materialises, operators carry the financing for distributed compute and backhaul, while suppliers of accelerated compute and RAN software hold the layer where incremental value would accumulate.
Disclosed checkpoints cluster in 2027: BT's Birmingham live trial, Deutsche Telekom's field trials, and Elisa's distributed deployment running commercial AI-RAN services.
