NVIDIA, Nokia and Operators Advance AI-RAN on Live Networks

NVIDIA announced on October 21, 2026, at GTC Berlin that global telecom operators are adding NVIDIA accelerated computing to centralized data centers, metro sites and cell sites, and that T-Mobile has completed what the company described as the world’s first AI-RAN call on a live commercial network.

“Telecom networks were built to carry the world’s conversations. Now, they will also carry the world’s intelligence,” said Jensen Huang, founder and CEO of NVIDIA. Huang said operators are turning their land, power and connectivity into distributed AI infrastructure, and that with Nokia, NVIDIA is extending accelerated computing into the radio network itself as a foundation for AI-native 5G.

NVIDIA frames intelligence as five layers — land, power and shell, chips, infrastructure, models and applications — and said telecom operators already hold sovereign infrastructure at the first layer: data centers, regional sites and cell sites with power, fiber and spectrum. By adding accelerated computing, open models and applications on top, the company said, operators can serve models and applications closer to the people, businesses and machines that use them, while a common converged platform based on the AI-RAN architecture at distributed sites can power robotics, voice and agentic applications and run intelligent 5G software.

Commercial AI-RAN Call and European Trials

AI-RAN extends accelerated computing into the radio access network, turning it into an intelligence workload that runs on the same platform as advanced services. NVIDIA said the architecture lets operators get more out of their spectrum while serving demand from sites close to users. Nearly 20 telecom operators globally have publicly announced AI-RAN collaborations using Nokia anyRAN software and NVIDIA Aerial RAN Computer (ARC), and Nokia is bringing its anyRAN software on ARC for trials at BT, Deutsche Telekom, Elisa and Orange.

T-Mobile’s call in Bellevue, Washington, ran on commercial ARC servers, Nokia’s NVIDIA CUDA-accelerated RAN software and massive multiple-input multiple-output radios, and connected to an existing 5G Standalone core and a commercial smartphone.

BT is planning live AI-RAN trials in Birmingham, England, with Nokia and NVIDIA in 2027. Deutsche Telekom has validated Nokia and NVIDIA’s commercial AI-RAN platform in Bonn and Oulu laboratories, including AI-enhanced uplink channel estimation intended to improve uplink user throughput, and plans to move to field trials in 2027.

Elisa simultaneously ran voice, mobile data and video analytics on the ARC platform in what NVIDIA described as a European first in Elisa’s over-the-air field environment using 5G spectrum, and NVIDIA said the trial clears the way for distributed deployment and commercial AI-RAN services in 2027. Orange is testing Nokia 5G software on ARC, focusing on AI-RAN spectral efficiency improvements and support for edge services in live networks.

Nine EMEA Operators Build NVIDIA-Powered Infrastructure

Nine operators across Europe, the Middle East and Africa are building or expanding NVIDIA-powered infrastructure across centralized and distributed locations. New infrastructure is being built by CRA in the Czech Republic, etisalat in the United Arab Emirates, Silknet in Georgia and TIM in Italy, while Cassava in Africa, Deutsche Telekom in Germany, Fastweb + Vodafone in Italy, Kazakhtelecom in Central Asia and Orion Telecom in Serbia are expanding existing infrastructure.

Deutsche Telekom plans to expand its NVIDIA-powered Industrial AI Cloud, which hosts open models including NVIDIA Nemotron and Soofi and serves robotics companies such as Agile Robots. A GTC Berlin demonstration with Agile Robots showed the approach: on-robot intelligence handles real-time perception and action, a model at the distributed edge supports fine-tuning and coordination across robots, and central infrastructure provides capacity for larger training workloads. At distributed sites, Deutsche Telekom is also piloting AI-RAN trials with Nokia and NVIDIA and demonstrating ARC running inference between the device and the network edge.

Orange is hosting open models, including Nemotron, as a managed service on its Live Intelligence trusted AI platform to support new enterprise services. At distributed sites, Orange is deploying ARC for voice intelligence applications using Nemotron Speech, demonstrating services that translate across multiple languages in real time for emergency and field services teams. Elisa has deployed ARC at a distributed network location, where it runs a voice intelligence application serving as an enterprise assistant, and is using open models to build edge applications.

New Aerial RAN Computer Configurations

NVIDIA introduced two ARC configurations, MGX 2U and MGX 4U, designed for dense, high-capacity intelligence and radio workloads typically deployed at centralized RAN sites. The configurations are powered by NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, and NVIDIA said they will offer x86 and NVIDIA Vera CPU options. Dell Technologies, Quanta Cloud Technology and Supermicro will offer ARC-ready servers, according to NVIDIA.

Nemotron Models Enter Network Operations Software

Operators and software partners are also adopting NVIDIA Nemotron open models and NVIDIA Agent Toolkit software, now integrated into commercial network software, to build agents for autonomous network operations. Nokia is integrating Nemotron into its Autonomous Networks Suite and using Agent Toolkit to build and deploy agents that manage operations across fixed, radio, core and transport networks.

Cisco has fine-tuned Nemotron models and integrated them into Crosswork AI, creating a common agentic layer across its Crosswork Network Automation software to manage IP and optical transport networks, and is using Agent Toolkit to help telcos operate agents safely. TCS is working with telcos to fine-tune Nemotron models and build specialized agents that validate radio-site builds, detect, diagnose and auto-remediate network issues, and deliver multilingual voice experiences.

BT is adopting Nemotron to build agents that inspect field-installation photos, detect problems and help reduce repeat site visits. NTT DATA is extending its NextGenOSS platform with fine-tuned Nemotron models and Agent Toolkit to build agents that detect early network degradation, identify likely root causes and recommend next actions.