Samsung Electronics and SoftBank Corp. officially signed a Memorandum of Understanding in October 2025 to collaborate on developing AI-based radio access network innovations and next-generation 6G communications.
The agreement underscores both companies’ intent to accelerate research in artificial intelligence and telecommunications synergy for global connectivity.
The partnership defines four candidate research pillars: 6G, AI for RAN, AI and RAN, and the Large Telecom Model.
Each field targets a distinct aspect of how artificial intelligence can enhance wireless network intelligence, reliability, and scalability as the telecom ecosystem moves toward sixth-generation infrastructure readiness.
What does the new AI-RAN collaboration aim to achieve?
AI-RAN aims to embed artificial intelligence directly within the radio access network to allow predictive operations, intelligent traffic handling, and self-optimization of base stations.
Samsung and SoftBank intend to establish architectures that enable distributed AI computing near baseband units, minimizing network latency and improving energy utilization within live telecom environments.
Both companies will conduct field trials to evaluate how AI interacts at the physical layer and in hybrid processing environments that combine central and edge computing.
Through these trials, they plan to demonstrate measurable gains in spectral efficiency and coverage performance. The results could lay the groundwork for early 6G network blueprints across Asia and North America.
Did you know?
6G networks are envisioned to deliver peak data rates of up to 1 terabit per second (Tbps), which is up to 1000 times faster than current 5G peak speeds, enabling truly immersive experiences like holographic communication.
How will 6G development benefit from this partnership?
The partnership highlights active discussions about the 7 GHz spectrum, one of the potential frequency bands under consideration for 6G use.
The centimeter-wave band provides a balance between high data throughput and manageable propagation loss, making it suitable for dense urban deployment and fixed wireless access scenarios.
Samsung’s advanced base-station technology, paired with SoftBank’s extensive deployment experience, will enable the exploration of realistic use cases.
Both parties aim to test mobile edge computing integration, distributed antenna arrays, and adaptive real-time processing frameworks that push 6G standardization beyond theoretical research.
Why is the Large Telecom Model crucial to network evolution?
The Large Telecom Model (LTM) represents SoftBank’s attempt to bring generative AI capabilities into network operations. Built on principles similar to large language models, LTM can transform network optimization by analyzing telemetry data, predicting anomalies, and proposing actions in a human-readable format.
This introduces a new layer of interpretability and automation to the network lifecycle. When integrated into existing RAN systems, LTM can optimize power allocation, schedule AI workloads efficiently, and even train sub-models for dynamic network adjustment.
Samsung’s engineering expertise complements this approach by ensuring that these AI-driven layers remain compliant with evolving global telecom protocols.
ALSO READ | Amazon trials driver glasses nationwide, aims to cut route times by 30 minutes
How will orchestration balance AI and RAN efficiency?
Samsung and SoftBank are jointly defining an AI-RAN Orchestrator that will intelligently schedule and prioritize workloads from both AI inference engines and RAN processing modules.
This orchestrator will ensure that AI computing does not interfere with real-time signal operations, but instead supports predictive decision-making.
Tests under the AI and RAN concept will examine how shared hardware infrastructure can host both computation domains seamlessly.
Efficient orchestration promises not only cost reduction but also sustainability improvements by consolidating AI and RAN on common hardware resources, reducing redundancy and energy waste.
What are the next milestones in AI-driven telecom research?
In November, Samsung will host the Silicon Valley Future Wireless Summit, bringing together academic, industrial, and policy leaders to shape international collaboration on AI-RAN standardization.
This event is expected to spark cross-sector dialogues that define the next phase of testbeds and interoperability frameworks for hybrid AI networks.
SoftBank continues developing models that bridge quantum computing, edge AI, and information theory for communications.
The joint research effort with Samsung hints at expanding beyond Japan and Korea, signaling pilot runs in cities across the United States and Europe once regulatory approvals are in place.
Together, Samsung and SoftBank have positioned this MoU as a step toward commercial AI-enabled telecom systems that can adapt autonomously and serve as the backbone for self-evolving 6G networks worldwide.
Their collaboration reinforces a shift toward an intelligent connectivity era where machine learning becomes a fundamental telecom function.
Comments (0)
Please sign in to leave a comment