Operators back AI‑native RAN
Nokia and Orange are working with Nvidia to test radio networks that run AI functions alongside traditional RAN tasks, moving AI‑RAN from vendor talk to operator-backed trials. They say the collaboration will identify and validate new AI‑RAN functions through a structured co‑innovation framework, signalling operators are helping set de‑facto architectures rather than waiting for standards bodies to decide. (za.investing.com; manilatimes.net)
Mobile operators are starting to test radio networks that handle artificial intelligence jobs and core wireless tasks on the same infrastructure, with Orange joining Nokia and Nvidia in a new April 15 effort. (nokia.com) A radio access network is the part of a mobile system that links phones to cell towers, and it normally runs tightly timed signal-processing work on specialized hardware. Nokia and Orange said they will evaluate artificial intelligence radio access network technology using Nokia anyRAN 5G software and Nvidia artificial intelligence infrastructure. (nokia.com) The companies said the project will test whether a graphics processing unit-based radio processor can improve receivers and whether artificial intelligence can be folded into functions such as scheduling, beamforming and power optimization. Orange said the work is aimed at better network performance, lower energy use and new services for customers. (nokia.com) Orange said the collaboration will run through a “structured co-innovation framework” in which the operator and Nokia jointly identify, design and evaluate new functions, rather than waiting for standards groups to settle every detail first. Orange’s Laurent Leboucher, the group chief technology officer, said the company wants to understand how those functions could fit into live networks across Europe, the Middle East and Africa. (nokia.com) That marks a shift from vendor demonstrations toward operator-backed trials. Nokia said on March 1 that it had already completed functional tests of graphics processing unit-accelerated artificial intelligence radio access network workloads with T-Mobile US, Indosat and SoftBank, and was also working with BT, Elisa, NTT Docomo and Vodafone Group. (nokia.com) The technical pitch is that one pool of computing power could run both wireless network functions and artificial intelligence applications, instead of keeping them on separate systems. Nokia and Orange said they will study whether that shared setup can raise spectral efficiency, improve predictive optimization and support sensing services alongside ordinary mobile traffic. (nokia.com) The spectrum piece matters because carriers are trying to squeeze more data out of existing airwaves before buying more. Nokia and Orange said the work will also cover current and future bands, including upper 6 gigahertz, as they prepare for a software-defined migration path toward sixth-generation, or 6G, networks. (nokia.com; telecoms.com) Competitors are making similar moves. Telecoms.com reported on April 15 that Ericsson and SK Telecom signed a memorandum last month to study how artificial intelligence radio access network techniques could improve performance, security and energy efficiency in fifth-generation, or 5G, networks. (telecoms.com) The open question is whether these trials become a repeatable commercial design instead of a lab exercise. Converge Digest reported on April 15 that operators are now testing whether artificial intelligence radio access network can justify itself on capacity and energy economics as the industry moves from 5G Advanced toward 6G. (convergedigest.com)
