Nvidia said on 21 October 2026, at GTC Berlin, that T-Mobile has completed “the world’s first AI-RAN call on a commercial network,” and that BT, Deutsche Telekom and Elisa have announced AI-RAN pilots on live networks using Nokia’s anyRAN software on Nvidia’s Aerial RAN Computer. Nvidia says nearly 20 telecom operators have now publicly announced AI-RAN collaborations using Nokia anyRAN software and Nvidia’s ARC.
In plain terms, the radio access network, or RAN, is the part of a mobile network that links phones to the cell towers and the equipment behind them. AI-RAN means running that radio software on the same accelerated computers that can also serve AI models, so a telco’s towers and local sites double as small data centres close to users. Nvidia calls the result a “distributed intelligence platform.”
Business Pill · INFERENCE
A Business Engineer explainer on inference, the cost paid on every AI answer. It covers the concept in general and says nothing about Nvidia, Nokia or the operators named.
Watch · The Keynote
NVIDIA GTC Berlin 2026 keynote with CEO Jensen Huang, 21 October, on NVIDIA’s YouTube channel. Quotations in this piece come from Nvidia’s release, not from the spoken keynote.
The key insight: As we read it, the release is less about a single product than about where Nvidia wants its computers to sit: in telco data centres, metro sites and cell sites, with Nokia’s radio software running on the same machines, and T-Mobile’s call on a commercial network offered as the first proof.
Nvidia’s release says leading operators are adding its accelerated computing “to centralized data centers, metro sites and cell sites so they can serve models and applications closer to the people, businesses and machines that use them.” It also says Nvidia is introducing new Aerial RAN Computer configurations for high-capacity sites.
“Telecom networks were built to carry the world’s conversations. Now, they will also carry the world’s intelligence,” Jensen Huang, Nvidia’s founder and CEO, said in the release. “With Nokia, we’re extending accelerated computing into the radio network itself, creating the foundation for AI-native 5G.”
Leading telecom operators are adding NVIDIA computing across data centers and cell sites to bring AI services closer to users, alongside live-network AI-RAN pilots with @nokia.
NVIDIA’s newsroom account posted the announcement at GTC Berlin, saying leading telecom operators are adding NVIDIA computing across data centers and cell sites to bring AI services closer to users, alongside live-network AI-RAN pilots with Nokia. The post links to the release this piece is built on.
Who Is Testing AI-RAN, and When
T-Mobile. The release says T-Mobile “recently marked” the first AI-RAN call on its commercial network in Bellevue, Washington, running on commercial ARC servers and Nokia’s CUDA-accelerated RAN software, connected to an existing 5G Standalone core and a commercial smartphone. It does not give the date of the call.
BT. Planning live AI-RAN trials in Birmingham, England, with Nokia and Nvidia in 2027, according to the release.
Deutsche Telekom. Has validated Nokia and Nvidia’s commercial AI-RAN platform in Bonn and Oulu labs, the release says, with plans to move to field trials in 2027.
Elisa. Ran voice, mobile data and video analytics at once on the ARC platform in an over-the-air field test using 5G spectrum, which Nvidia calls a European first; the release says the trial “clears the way for distributed deployment and commercial AI-RAN services in 2027.”
Orange. Testing Nokia 5G software on ARC, focused on spectral efficiency and edge services in live networks.
Telcos NVIDIA’s 21 October release names: five with AI-RAN trials or calls (T-Mobile, BT, Deutsche Telekom, Elisa, Orange), four new EMEA builds and five EMEA expansions. The counts are our arithmetic; “nearly 20” telcos with public AI-RAN collaborations is Nvidia’s figure. “Three trials point to 2027” refers to BT, Deutsche Telekom and Elisa. Source: NVIDIA release, GTC Berlin, 21 October 2026.
Europe’s Telcos Add Nvidia Computing
Beyond the radio network, the release says nine EMEA operators are building or expanding Nvidia-powered infrastructure. New builds: CRA in the Czech Republic, etisalat in the United Arab Emirates, Silknet in Georgia and TIM in Italy. Expansions: Cassava in Africa, Deutsche Telekom in Germany, Fastweb + Vodafone in Italy, Kazakhtelecom in Central Asia and Orion Telecom in Serbia.
Deutsche Telekom plans to expand its Nvidia-powered Industrial AI Cloud, which the release says hosts open models including Nvidia Nemotron and Soofi and serves robotics companies such as Agile Robots. Orange is hosting open models, including Nemotron, as a managed service on its Live Intelligence platform, and Elisa has deployed Nvidia ARC at a distributed network location to run a voice assistant for enterprises.
Agents to Run the Network
The release names four software partners building agents for telecom operations on Nvidia’s Nemotron open models and Agent Toolkit. Nokia is integrating Nemotron into its Autonomous Networks Suite; Cisco has fine-tuned Nemotron models for its Crosswork AI network automation software; TCS is working with telcos on agents that validate radio-site builds and diagnose network issues; and NTT DATA is extending its NextGenOSS platform. BT, the release says, is adopting Nemotron to build agents that inspect field-installation photos and help reduce repeat site visits.
What Huang Said on Stage
In the keynote itself, Huang went further than the release. He predicted that every telco, which he put at roughly 100,000 sites and around 100 gigawatts of distributed base stations and data centres, will in future evolve into a distributed AI network, according to a transcript of Nvidia’s keynote stream. That is his forecast; the release does not repeat the 100-gigawatt figure.
Connecting the Coverage
This is the fifth GTC Berlin release we have covered today. Our DSX water piece reported Nvidia’s claim that its AI factory designs can cut cooling water use to near zero, and our DSX POD piece reported its claim that factory-built units cut deployment at a prepared site to as little as 14 days.
Our OpenShell piece covered Nvidia’s blueprint for running AI agents under enforced limits, and our open-models piece reported that SAP, Salesforce and others are building agents on Nemotron. What is new here is where Nvidia wants that computing to sit: not only in large AI factories but in telco sites close to users, with Nokia supplying the radio software.
The Structural Read
As we read it, the radio network has been a separate world of specialised equipment; the release describes AI-RAN as turning it into “an intelligence workload that runs on the same platform as advanced services,” which would put Nvidia hardware inside the telco equipment market.
As we read it, the dates matter as much as the names: BT plans live trials in 2027, Deutsche Telekom plans field trials in 2027 and Elisa points to commercial AI-RAN services in 2027, so most of the radio-network work Nvidia names is still ahead.
As we read it, the European infrastructure list is the nearer-term business: nine EMEA operators building or expanding Nvidia-powered computing, several of them hosting open models such as Nemotron as a service for their own enterprise customers.
Jensen Huang, founder and CEO of NVIDIA, in Nvidia’s release of 21 October 2026
“Leading global operators are turning their land, power and connectivity into a distributed AI infrastructure — bringing intelligence closer to the industries, people and robots that use it.”
Three Implications
FOR TELECOM OPERATORS Nvidia says nearly 20 operators have publicly announced AI-RAN collaborations using Nokia anyRAN software and Nvidia ARC, with BT, Deutsche Telekom and Elisa naming 2027 for trials or services.
FOR ENTERPRISES THAT BUY FROM TELCOS The release says Orange hosts open models including Nemotron as a managed service, and that Deutsche Telekom’s Industrial AI Cloud serves robotics companies such as Agile Robots.
FOR TELECOM SOFTWARE VENDORS Nokia, Cisco, TCS and NTT DATA are named as building network-operations agents on Nemotron and Nvidia’s Agent Toolkit.
The analysis contrasts training with inference: “Inference is a revenue engine: continuous, distributed across millions of applications and agents, incurring marginal cost with every query.”
As we read it, Nvidia’s telecom release is a bet on that distribution: if inference runs continuously and everywhere, the sites closest to users, which telcos already own, become places to sell computing. The analysis was published in March 2026 and does not discuss AI-RAN or these operators.
What Is Not Established
The release, as we read it on 21 October, gives no prices, contract values, capacity figures or numbers of sites for any operator, and does not say when T-Mobile’s call took place. Most of the named radio-network work is still in trials, with BT, Deutsche Telekom and Elisa pointing to 2027. The counts of named telcos in our poster are our arithmetic; the “nearly 20” figure is Nvidia’s. Huang’s 100-gigawatt figure is a stage forecast, not a disclosed plan.
Business Engineer Framework
The State of The Inference Economy
A Business Engineer analysis of inference as a continuous, distributed cost and revenue engine for AI.
Nvidia used GTC Berlin to say T-Mobile has made the first AI-RAN call on a commercial network, and to name BT, Deutsche Telekom, Elisa and Orange among the operators testing Nokia’s radio software on its computers. As we read it, the nearer-term business is the nine EMEA telcos adding Nvidia computing and hosting open models, while most of the AI-RAN trials and services the release names point to 2027.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. This piece is built on NVIDIA’s press release of 21 October 2026 on telecom networks and AI-RAN, read in full; quotations come from that release only. Huang’s stage forecast is paraphrased from a transcript of Nvidia’s keynote stream. The counts of named telcos are our arithmetic. Nokia had not published its own release on Nokia.com when we checked. We did not contact Nvidia, Nokia or the operators named. Nothing here is legal, financial or investment advice.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
