Nokia — at MWC 2025 — announced its partnership with KDDI, SoftBank, T-Mobile US, and NVIDIA to advance AI-powered Radio Access Networks (AI-RAN).

Nokia will be establishing AI-RAN center in Dallas for real-world testing, innovation, and validation of AI-RAN solutions.
Nokia has integrated AI computing into shared infrastructure for better efficiency.
AI-RAN enables new business models via Platform-as-a-Service (PaaS) offerings.
Nokia’s partnership with NVIDIA, SoftBank, and T-Mobile is aimed to maximize AI-RAN efficiency and feasibility.
Nokia will be showcasing AI-powered AirScale base stations and MantaRay AutoPilot for autonomous RAN operations.
Nokia’s AI-RAN for KDDI
Nokia and KDDI are collaborating to explore AI-RAN use cases and architectures, aiming for commercial viability. Their focus is on enhancing user experience, improving network quality, reducing operational costs, and optimizing power consumption through AI-driven automation. The partnership will conduct commercial trials with AI-enabled RAN hardware and investigate AI-driven network performance enhancements.
“We anticipate that AI-RAN will unlock significant network optimization, enhance user experiences, reduce costs, and generating new services and revenue, leading to a more efficient and intelligent 5G ecosystem,” said Kazuhiro Furuhata, Executive Officer & General Manager, Network Node Technical Development Division Core Technology Sector at KDDI.
Nokia’s AI-RAN for SoftBank
Nokia and SoftBank have integrated AI workloads into an AI-RAN platform using Red Hat OpenShift, managed through Nokia’s MantaRay NM and SoftBank’s AITRAS Orchestrator. This collaboration demonstrates how RAN and non-RAN AI workloads can share computing resources efficiently, improving operational efficiencies and accelerating ROI for network operators.
“Through the monitoring of hardware resources by the AITRAS Orchestrator, we have successfully enabled the coexistence of vRAN and AI applications. This advancement facilitates the more efficient utilization of base station equipment,” said Hideyuki Tsukuda, Executive Vice President & CTO, SoftBank
Nokia’s AI-RAN for T-Mobile US
Nokia and T-Mobile US are working on multi-purpose cloud infrastructure for AI-RAN. They are evaluating AI-RAN architecture, testing the feasibility of accelerated computing for Layer 1 (L1), and exploring how AI and RAN can coexist on shared infrastructure using Nokia Cloud RAN and NVIDIA platforms. The partnership is also investigating monetization opportunities and the broader economics of AI-RAN.
“By exploring AI-driven architectures and leveraging multi-purpose cloud infrastructure, we’re evaluating how accelerated compute for Layer 1 (L1) and the integration of AI and RAN on shared platforms with our industry partners will enhance network performance and efficiency,” said John Saw, Executive Vice President, Chief Technology Officer, T-Mobile.
Baburajan Kizhakedath