Finland's Elisa announced a partnership with NVIDIA and Nokia to validate AI‑RAN in a live commercial network slated for 2027. The collaboration builds on Europe’s first AI‑RAN trials, showcasing voice, data and video analytics on a shared 5G‑Advanced platform.
Elisa Oyj disclosed on Oct. 21, 2026 that it is moving toward commercial AI‑RAN (Artificial Intelligence‑Radio Access Network) by deploying distributed AI infrastructure across its Finnish 5G‑Advanced network. The operator will run a live commercial trial in 2027 with technology supplied by NVIDIA and Nokia, following Europe’s first live‑network AI‑RAN trials completed earlier this year (source [2]).
The AI‑RAN platform merges radio access functions and AI workloads on the same accelerated computing hardware, allowing the network to process AI inference close to the data source. In a recent outdoor test, Elisa simultaneously ran voice calls, mobile data and video‑analytics workloads on the same AI‑RAN platform, demonstrating the feasibility of real‑time, edge‑based AI services (source [2]).
Key hardware components include Nokia’s AI‑RAN solution paired with NVIDIA’s Aerial RAN Computer (ARC) equipped with RTX PRO 4500 Blackwell Server Edition GPUs, as well as Dell PowerEdge servers fitted with RTX PRO 6000 BSE GPUs for broader deployment (source [2]).
"AI is becoming a critical layer of digital infrastructure," said Sami Komulainen, Elisa’s COO and EVP of Technology and Operations. "Our goal is to bring AI closer to users, enterprises and data by advancing a distributed AI inference network in Finland," he added (source [2]).
Representatives from the partners echoed the significance of the collaboration. Ronnie Vasishta of NVIDIA highlighted the shift of AI from data centres to the network edge, while Emma Falck of Nokia emphasized that AI‑RAN will enable smarter networks, more efficient infrastructure use and new revenue streams from edge AI services (source [2]).
Elisa’s roadmap aims to validate commercial benefits in 2027, paving the way for AI‑native 6G networks and a new class of low‑latency, AI‑driven applications such as instant voice assistants, on‑site video safety analytics and autonomous network optimization (source [2]).