SoftBank’s AITRAS orchestration, integrated with Ericsson’s O-RAN-compliant SMO, Nokia’s AI platform, and the OCUDU ecosystem, is designed to dynamically allocate computing resources across AI and RAN domains, fostering an open global AI-RAN ecosystem.

At MWC Barcelona 2026, SoftBank announced a series of moves: a joint PoC with Ericsson for low-latency, high-reliability AI-RAN-enabled networks; a secure edge AI infrastructure deployment with Mitsubishi Heavy Industries for on-premises AI inference in industrial sectors; and an expansion of the AI-RAN ecosystem through open-sourcing the Dynamic Scoring Framework and broader OSS collaboration.

SoftBank demonstrated AI-RAN capabilities in the Ericsson PoC, showcasing tight integration of robots, networks, and computing for real-time AI offloading and control with low latency.

Contextual notes reference related press releases, including a January 21, 2026 note about Infrinia AI Cloud OS, reinforcing the broader strategy and timeline.

Market context suggests AI-related SoftBank/ERIC news historically yields modest single-digit stock moves, indicating tempered investor expectations for AI-infrastructure initiatives.

Background includes prior collaboration with Yaskawa Electric on Physical AI robots and earlier SoftBank announcements on Infrinia AI Cloud OS and AITRAS orchestration.

The architecture links large data centers across Japan with nationwide distributed GPU resources, edge Aitrus devices, and regional brain clusters to support AI model training and inference while addressing latency through edge GPUs and centralized training.

SoftBank Research Institute of Advanced Technology leads R&D and industry-academia collaboration to advance AI-RAN, Beyond 5G/6G, and related technologies.

An on-premises field trial in a factory setting with Mitsubishi Heavy Industries demonstrated enterprise applicability of AI-RAN using Aitras, with stable 5G connectivity and nearby GPUs.

Related press releases underscore ongoing collaborations with Ericsson, Mitsubishi Heavy Industries, and open-source community efforts.

SoftBank unveils Telco AI Cloud as a vision to transform telecom networks into AI infrastructure, leveraging GPU clouds, AI-RAN MEC, and Infrinia AI Cloud OS for distributed, low-latency processing with data sovereignty.

Infrinia AI Cloud OS tackles multi-site operational complexity by integrating management of GPUs, telecom networks, Kubernetes, and AI workloads for optimized training and inference in a multi-tenant environment.