Researchers Demo Agentic AI for Radio Access Networks

Published February 26, 2026 by The Daily Scout

Northeastern University, SoftBank, Keysight, and zTouch Networks demonstrated an autonomous, agentic AI-RAN at MWC Barcelona 2026. The system is powered by a Large Telecom Model (LTM) and is designed to be an intent-driven, AI-native Radio Access Network.

Why it matters

- The "Large Telecom Model" (LTM) is a generative AI foundation model trained on vast datasets from SoftBank's network, designed to understand and execute complex network operations from high-level natural language instructions. This approach aims to reduce workflow times from days to minutes, a significant decrease in operational time and cost. - An AI-Radio Access Network (AI-RAN) integrates AI deeply into the mobile network's radio access layer to enhance performance, improve spectral efficiency, and automate management. This technology represents a strategic shift from simply automating tasks to creating a fully autonomous, intelligent network fabric capable of self-optimization and healing. - The demonstration at MWC Barcelona 2026 is a product of the AI-RAN Alliance, a consortium that includes founding members like NVIDIA, Amazon Web Services, Ericsson, Nokia, and T-Mobile, indicating a broad industry push toward AI-native wireless networks. This collaboration focuses on integrating AI to improve network efficiency, lower power consumption, and open new revenue streams as 5G and 6G technologies evolve. - For engineers in NYC, the rise of agentic AI is creating a demand for new skills, with local companies like Assembled, Eulerity, and Hunch actively hiring for roles in AI and automation. The city's tech scene includes a growing number of startups focused on vertical AI applications, from AI-powered copilots for psychiatry at Blossom to AI agents for power projects at Paces. - NYC's venture capital landscape is actively funding AI-native startups. Insight Partners and Thrive Capital are notable late-stage investors, while firms like First Round Capital focus on seed-stage companies. Recently, AI accounting startup Basis, headquartered in NYC, raised $100 million in a Series B round, demonstrating significant investor confidence in AI agent applications for specific industries. - Engineers looking to build on the side can draw inspiration from indie hackers who have successfully transitioned from employee to entrepreneur. One developer, after growing a Twitter following, generated $577,000 in a week-long presale for a CSS course. Another software engineer in NYC detailed earning an additional $2,000 a month by building AI agents for small businesses in just 6-8 hours of work per month. - Practical blueprints for transitioning from enterprise to startup often involve starting with a side project to gain traction before leaving a full-time role. One founder launched their first successful side project in 2019 while still employed, eventually leaving their Head of Software role in 2024 to focus on a new venture full-time. This path allows for product validation and revenue generation, de-risking the leap into full-time entrepreneurship. - The development of agentic AI is not limited to network optimization; it's also being applied to create entirely new services. At MWC Barcelona 2026, companies like Nokia are showcasing how AI agents can enable on-demand network slicing for premium 5G services at mass events or for emergency first responders. This highlights a move from using AI as a cost-saving tool to a revenue-generating one.

Key numbers

Northeastern University, SoftBank, Keysight, and zTouch Networks demonstrated an autonomous, agentic AI-RAN at MWC Barcelona 2026.
The demonstration at MWC Barcelona 2026 is a product of the AI-RAN Alliance, a consortium that includes founding members like NVIDIA, Amazon Web Services, Ericsson, Nokia, and T-Mobile, indicating a broad industry push toward AI-native wireless networks.
This collaboration focuses on integrating AI to improve network efficiency, lower power consumption, and open new revenue streams as 5G and 6G technologies evolve.
Recently, AI accounting startup Basis, headquartered in NYC, raised $100 million in a Series B round, demonstrating significant investor confidence in AI agent applications for specific industries.

What happens next

This approach aims to reduce workflow times from days to minutes, a significant decrease in operational time and cost.

Sources

Quick answers

What happened in Researchers Demo Agentic AI for Radio Access Networks?

Why does Researchers Demo Agentic AI for Radio Access Networks matter?

The "Large Telecom Model" (LTM) is a generative AI foundation model trained on vast datasets from SoftBank's network, designed to understand and execute complex network operations from high-level natural language instructions. This approach aims to reduce workflow times from days to minutes, a significant decrease in operational time and cost. An AI-Radio Access Network (AI-RAN) integrates AI deeply into the mobile network's radio access layer to enhance performance, improve spectral efficiency, and automate management. This technology represents a strategic shift from simply automating tasks to creating a fully autonomous, intelligent network fabric capable of self-optimization and healing. The demonstration at MWC Barcelona 2026 is a product of the AI-RAN Alliance, a consortium that includes founding members like NVIDIA, Amazon Web Services, Ericsson, Nokia, and T-Mobile, indicating a broad industry push toward AI-native wireless networks. This collaboration focuses on integrating AI to improve network efficiency, lower power consumption, and open new revenue streams as 5G and 6G technologies evolve. For engineers in NYC, the rise of agentic AI is creating a demand for new skills, with local companies like Assembled, Eulerity, and Hunch actively hiring for roles in AI and automation. The city's tech scene includes a growing number of startups focused on vertical AI applications, from AI-powered copilots for psychiatry at Blossom to AI agents for power projects at Paces. NYC's venture capital landscape is actively funding AI-native startups. Insight Partners and Thrive Capital are notable late-stage investors, while firms like First Round Capital focus on seed-stage companies. Recently, AI accounting startup Basis, headquartered in NYC, raised $100 million in a Series B round, demonstrating significant investor confidence in AI agent applications for specific industries. Engineers looking to build on the side can draw inspiration from indie hackers who have successfully transitioned from employee to entrepreneur. One developer, after growing a Twitter following, generated $577,000 in a week-long presale for a CSS course. Another software engineer in NYC detailed earning an additional $2,000 a month by building AI agents for small businesses in just 6-8 hours of work per month. Practical blueprints for transitioning from enterprise to startup often involve starting with a side project to gain traction before leaving a full-time role. One founder launched their first successful side project in 2019 while still employed, eventually leaving their Head of Software role in 2024 to focus on a new venture full-time. This path allows for product validation and revenue generation, de-risking the leap into full-time entrepreneurship. The development of agentic AI is not limited to network optimization; it's also being applied to create entirely new services. At MWC Barcelona 2026, companies like Nokia are showcasing how AI agents can enable on-demand network slicing for premium 5G services at mass events or for emergency first responders. This highlights a move from using AI as a cost-saving tool to a revenue-generating one.