T-Mobile completes live-network AI-RAN call with Nokia and NVIDIA
The Bellevue call used commercial network equipment and a smartphone. Follow-up trials will explore AI workloads sharing infrastructure with radio functions.
Listen to this story
The audio brief
Story brief
3 key pointsOn October 21, 2026, Nokia, T-Mobile and NVIDIA completed a call using AI-RAN equipment at a Bellevue cell site connected to T-Mobile’s commercial network. The demonstration moves the radio setup from lab testing into a live operator environment, but it does not show AI inference running alongside radio functions or establish performance gains. Planned field trials will test shared AI and network workloads, giving operators and suppliers a chance to assess whether the architecture can work beyond a single call.
- 01
The setup paired Nokia Massive MIMO radios and anyRAN software with servers using NVIDIA’s Aerial RAN Computer platform.
- 02
NVIDIA said the call used an existing commercial 5G Standalone core and a commercial smartphone.
- 03
Nokia’s goal of doubling spectral efficiency remains an ambition, not a result measured in the Bellevue call.
AI-RAN has moved from laboratory testing into T-Mobile’s live commercial network. On October 21, 2026, Nokia said it, T-Mobile and NVIDIA completed a call at a commercial cell site in Bellevue, Washington. The milestone puts their accelerated-computing approach into an operating mobile network; trials of AI workloads sharing that infrastructure are still ahead.
A commercial phone, core network and cell site
The call combined Nokia’s Massive MIMO radio equipment and 5G anyRAN software with servers built on NVIDIA’s Aerial RAN Computer platform. NVIDIA’s announcement added that the setup connected to an existing commercial 5G Standalone core and a commercial smartphone. Those connections placed the call within T-Mobile’s production network, rather than keeping the entire setup inside a laboratory.
AI-RAN brings accelerated computing into the radio access network, or RAN. The companies’ intended architecture uses a common computing platform for radio functions and AI services. The Bellevue call demonstrated the radio-network side of that effort on commercial infrastructure. Running AI inference—the work of using a trained model—alongside those radio functions is part of the next phase of field testing.
Nokia described the achievement as both the world’s first AI-RAN call and the first software-defined RAN call using accelerated computing. NVIDIA also called it a world first on a commercial network. The world-first claims come from the participating suppliers.
Bellevue follows a staged development effort
The call follows several steps in the same collaboration. T-Mobile announced an AI-RAN Innovation Center in Bellevue in September 2024. In 2025, Nokia and NVIDIA formed a partnership to bring NVIDIA-powered, commercial-grade AI-RAN products into Nokia’s radio-access portfolio. T-Mobile was also working with the two suppliers to integrate AI-RAN technologies into its 6G development process.
Earlier in 2026, Nokia announced successful laboratory testing of its anyRAN software on NVIDIA’s GPU-accelerated AI-RAN platform with T-Mobile and other operators. The Bellevue work advances that testing into a live operator environment with real spectrum. Nokia says the preparation involved months of engineering, including site preparation, integration, configuration, testing and optimization before the system was brought into operation.
The next trials add work beyond the call
The companies plan to extend field trials to additional spectrum bands, new AI-for-RAN algorithms and AI workloads running alongside radio-access functions on shared accelerated computing infrastructure. The planned tests therefore go beyond making another call: they will explore how the platform supports both network functions and AI processing in a field setting.
Nokia’s proposed business case joins two objectives. It wants better radio performance and more efficient use of network resources, while giving operators opportunities to earn revenue from AI services. In that vision, connectivity and AI computing would share infrastructure rather than remain separate activities. Nokia presents those benefits as potential outcomes, with further field deployments needed to understand how AI-RAN could operate at scale.
Sources
- globenewswire.comNVIDIA, Nokia and Leading Telecom Operators Turn Networks Into Distributed Intelligence Platforms
- nokia.comNokia, T-Mobile, NVIDIA complete world-first AI-RAN call
Reader comments
Newest comments first. Replies stay oldest first.