Nokia and NVIDIA advance AI-RAN with European operator trials

Nokia and NVIDIA have deepened their AI-RAN partnership, adding agentic network automation and signing up BT, Deutsche Telekom, Elisa and Orange for trials.

By The Datatech Times Editorial Team AI level 2 of 5
AI level 2 of 5: drafted by our AI editorial assistant from the company's announcement; fact-checked, edited and signed off by Mark Walker, Editorial Director. What the levels mean
A long, brightly lit data center aisle features rows of white server racks housing dark IT equipment, organized network cables, and overhead cable trays.

Nokia and NVIDIA have announced a set of advances to their joint AI-RAN programme, moving from platform definition toward live-network integration across four major European operators and launching a sovereign agentic AI platform aimed at cross-domain network autonomy.

The two companies first announced a commercial AI-RAN partnership in July, combining Nokia's AI-native anyRAN software with NVIDIA's Aerial RAN Computer, a GPU-accelerated platform for running AI workloads inside the radio access network. The October update extends that work into signal-processing functions including channel estimation, beamforming, multi-user MIMO pairing and scheduling, and adds a new agentic layer built on NVIDIA AI Enterprise software and Nemotron models.

Operator trials gather pace

Four European carriers are now engaged in varying stages of testing. Elisa has already run voice, mobile data and video analytics simultaneously on an AI-RAN platform in its live Helsinki 5G network, with distributed commercial deployment and customer use-case evaluation expected in 2027. BT is preparing a live-network trial in Birmingham, also targeting 2027. Deutsche Telekom and Nokia have reported initial laboratory measurements showing uplink throughput improvements of 10 to 15 per cent from an AI-based channel-estimation technique, with field validation and additional spectral-efficiency algorithms as the next step. Orange plans lab and live-network testing but has not published a timeline.

Pallavi Mahajan, Chief Technology and AI Officer at Nokia, said the partnership was "turning this vision into reality by bringing AI into the radio network and across network operations," describing the goal as a trusted software-defined foundation that helps operators extract more capacity from existing spectrum while opening a path toward AI-native 6G.

The operator line-up also includes etisalat UAE, MEO and Airtel, which are working with Nokia on the agentic platform to identify use cases and integration timelines, though no deployment dates have been confirmed for those carriers.

Agentic platform and ecosystem build-out

Beyond radio processing, Nokia and NVIDIA have co-developed what they describe as a sovereign agentic AI platform for network autonomy. The architecture allows operators to deploy telco-specific AI agents across radio, core, transport and cloud domains, with governance controls Nokia calls "glass-box autonomy." Operational data, models and policies can be kept on-premises, in the cloud or in a hybrid arrangement, which the companies position as meeting data-residency concerns for carriers operating under national regulatory frameworks.

The platform arrives at a point when network operators are under pressure from regulators in Europe and the Gulf to maintain control over sensitive network data. The EU's NIS2 Directive and individual national security requirements in the UK and Germany both impose obligations on carriers around critical infrastructure, and sovereign deployment options are increasingly a procurement requirement rather than a differentiator.

Nokia is also building out an edge-application ecosystem. DeepSig is developing spectrum-sensing software using existing radio infrastructure for RF threat detection, while ZaiNar is integrating precise network-based positioning capabilities into Nokia Cloud AI-RAN without requiring device-side changes. Infrastructure partners including Dell Technologies, Quanta Cloud Technology, Supermicro and Red Hat are supporting flexible deployment configurations and a cloud-native platform built on Red Hat OpenShift.

Market context

The AI-RAN category is still nascent. Nokia's principal competitors in the RAN market, Ericsson and Samsung, are pursuing their own AI-assisted radio software programmes, while a number of Open RAN software vendors are positioning AI-driven processing as a differentiator in virtualised architectures. Hyperscalers including Microsoft and Google have signalled interest in the telco edge as a deployment surface for AI inference workloads, which may position the RAN itself as contested territory between network-equipment vendors and cloud providers.

Nokia also highlighted 800G coherent optical and IP gateway solutions at GTC Berlin as the connectivity fabric needed to link distributed AI infrastructure across radio, edge and data-centre tiers. The near-term milestones to watch are BT's Birmingham trial results and Deutsche Telekom's field validation of the channel-estimation throughput gains reported in the laboratory.

Sector

Topic

Geography