Northeastern University, SoftBank Corp., Keysight, and zTouch Networks Demonstrate LTM-powered Autonomous Agentic AI-RAN at MWC Barcelona 2026

Long-standing industry-academic collaboration delivers first-of-its-kind intent-driven, AI-native RAN powered by a Large Telecom Model

BOSTON and BARCELONA, Spain, Feb. 26, 2026 /PRNewswire/ -- Northeastern University's Institute for Intelligent Networked Systems (INSI), SoftBank Corp. (SoftBank), Keysight Technologies, and zTouch Networks are demonstrating Autonomous Agentic AI-RAN (AgentRAN) at Mobile World Congress (MWC) Barcelona 2026. AgentRAN is a first-of-its-kind system that coordinates a hierarchy of AI agents to translate high-level operator intents into real-time, autonomous 5G and 6G network configurations and services.

The demonstration is the latest milestone in a collaboration anchored at Northeastern University, where INSI has served as a nexus for industry-academic partnerships at the frontier of intelligent wireless networks. Over the past few years, Northeastern has built deep research ties with SoftBank on AI-native architectures, leading to multiple publications, AI-RAN Alliance demonstrations, and innovation in next-generation wireless. INSI collaborates with Keysight on AI-RAN testing and spun off zTouch Networks to commercialize technology on open, GPU-accelerated RAN infrastructure. This demo brings those threads together for the first time around a shared vision: truly autonomous, AI-driven radio access networks.

How It Works

At the heart of the demo is AgentRAN, developed at Northeastern's Open6G AI-RAN Alliance Lab. Operators express goals in natural language – such as "maximize throughput while prioritizing emergency traffic" – and the AgentRAN Manager decomposes these intents into actions for specialized agents operating across the RAN stack. The agents are powered by the SoftBank's Large Telecom Model (LTM), a purpose-built foundation model trained on telecom-specific KPIs, configurations, and domain data—further optimized with high-fidelity training data generated using Keysight's RF digital twin channel emulation solutions with real world accurate 3D ray tracing capabilities (RaySIM, PROPSIM) for more accurate decisions than generic cloud models can deliver.

The Northeastern Open6G AI-RAN lab, using Keysight's PROPSIM, demonstrates multiple agentic scenarios, deployed on an end-to-end programmable private 5G AI-RAN infrastructure, managed and orchestrated by zTouch Networks' zTouch.OS. The demonstration shows the system's ability to detect changing conditions and adapt autonomously. Using Keysight's AI-RAN performance scoring solution, the system shows clear performance gains linked to the use of SoftBank's LTM, high fidelity training datasets generated by Keysight RF digital twin solutions, and zTouch Networks' zTouch.OS.

What the Partners Are Saying

Tommaso Melodia, William L. Smith Professor at Northeastern University and INSI Director: "When Northeastern co-founded the AI-RAN Alliance alongside SoftBank, NVIDIA, and other industry leaders at MWC 2024, we set out to prove that AI can be woven directly into the fabric of the radio access network – not as an afterthought, but as its core operating principle. This demonstration is proof of this concept coming to life. Our open AI-RAN research infrastructure at INSI was purpose-built to bring together partners like SoftBank, Keysight, and zTouch Networks so we can move from ideas to working, validated systems faster than any one organization could alone. What you see here is a network that understands what an operator wants and figures out how to deliver it autonomously. This is the future of wireless, and it is being built through exactly this kind of deep, sustained collaboration."

Ryuji Wakikawa, Vice President and Head of the Research Institute of Advanced Technology at SoftBank Corp.: "This initiative demonstrates our Large Telecom Model working outside of our own infrastructure for the first time, which underscores the potential of this technology to transform the whole industry. We believe technology is only valuable when it is put to the service of society, and our team is glad to support the spread of this impact in other use-cases beyond our own."

Peng Cao, Vice President and General Manager, Wireless Test Group, Keysight: "AI-native networks will only be as powerful as the data and validation frameworks behind them. Through this collaboration with Northeastern University, SoftBank, and zTouch Networks, Keysight is enabling the high-fidelity digital twin environments and performance intelligence needed to train, validate, and scale Large Telecom Models for real-world deployment. Together, we are helping the industry move forward with confidence toward an autonomous, trustworthy, AI-powered 6G era."

Salvatore D'Oro, zTouch Networks CTO: "We built zTouch Networks on a simple conviction: networks should understand what you need and deliver it autonomously - no armies of engineers, no months-long deployment cycles. This demonstration brings that vision to life. With zTouch.OS orchestrating an end-to-end AI-RAN infrastructure, operators can express an intent in plain language and watch the network configure itself in real time. Together with Northeastern, SoftBank, and Keysight, we're showing that truly autonomous, intent-driven wireless networks aren't a future promise - they're here."

The Agentic AI-RAN demonstration will be live at MWC Barcelona 2026, March 2-5, at the Keysight booth in Hall 5, Stand 5F41. Visitors can see the full intent-to-network pipeline in action and explore the AgentRAN dashboard, LTM integration, and real-time channel emulation firsthand.

About the Northeastern University INSI

The Institute for Intelligent Networked Systems (INSI) at Northeastern University conducts cutting-edge research in wireless communications, networking, and systems. Home to the Colosseum wireless network emulator and the Open6G R&D center, INSI leads multiple federally, industry, and state-funded programs advancing wireless technology.

Contacts: Michele Polese [email protected]

SOURCE Institute for Intelligent Networked Systems