Mobile World Congress (MWC) 2026 marked a turning point for AI-native networks. What had been a recurring theme in previous years materialized into announcements, products, and collaborations.

Nvidia and the Global Coalition for AI-RAN and 6G

Nvidia has struck deals with operators and technology companies, including BT Group, Deutsche Telekom, Ericsson, Nokia, SK Telecom, SoftBank, and T-Mobile, to develop 6G on open, secure, and AI-native platforms. The initiative, supported by collaborations with governments in the United States, United Kingdom, Europe, Japan, and Korea, aims to ensure intelligent and reliable connectivity infrastructures.

Nvidia has released open-source tools for network operators, including the Nemotron Large Telco Model (LTM) and guides for creating AI agents. The company also presented Nvidia Blueprints for RAN energy efficiency and network configuration. Cassava Technologies is implementing the latter for an autonomous network platform in Africa, while NTT DATA is using it in Japan to manage traffic surges.

Nokia and Operators Take AI-RAN to the Field

Nokia has made progress in its strategic AI-RAN partnership with Nvidia, completing tests of the anyRAN software on Nvidia's GPU-accelerated AI-RAN platform with T-Mobile US, Indosat Ooredoo Hutchison (IOH), and SoftBank Corp. T-Mobile ran AI and RAN workloads on a single Nvidia Grace Hopper 200 server. IOH made the first AI-RAN-powered Layer 3 5G call in Southeast Asia. SoftBank demonstrated how spare compute capacity can run third-party AI workloads.

Nokia's AI-RAN ecosystem now includes Dell Technologies, Quanta, Supermicro, and Red Hat OpenShift for orchestration.

Ericsson Takes a Different Approach

Ericsson unveiled ten new AI-ready radios based on proprietary silicio, with neural network accelerators integrated directly into the Massive MIMO hardware. According to Ericsson, this approach offers better TCO and energy efficiency than external GPU hardware. Ericsson also announced a collaboration with Intel to accelerate ecosystem readiness for AI-native 6G.

SK Telecom, SoftBank, and the Operator Rebuild

SK Telecom outlined a complete AI-native rebuild, from the network core to customer service systems, with plans to upgrade the sovereign AI foundation model and build a new AI data center in Korea. SoftBank demonstrated the Autonomous Agentic AI-RAN (AgentRAN) system that translates operator goals into real-time 5G and 6G network configurations.

A Hardware Ecosystem Takes Shape

Quanta Cloud Technology announced commercial AI-RAN products supporting Nvidia ARC platforms and Nokia software. Supermicro has extended support to the entire Nvidia AI-RAN portfolio. MSI introduced the unified AI-vRAN platform with dynamic GPU allocation. Lanner Electronics launched the AstraEdge AI servers, designed for the co-location of AI inference, RAN functions, and high-performance packet processing at cell sites. AMD positioned the EPYC 8005 edge platform and the Open Telco AI initiative as an alternative compute path for operators.

AI-RAN networks that continuously evolve through software mean that connectivity infrastructure increasingly resembles cloud infrastructure in terms of speed of change and flexibility. The integration of GPU computing within the RAN opens the prospect of running enterprise AI workloads at the network edge.