GSMA | AI Telco Troubleshooting Challenge Launched to Transform Network Reliability

Supported by GSMA, ETSI, IEEE GenAINet, ITU and TM Forum, the new initiative invites telecom innovators to develop LLM for root cause analysis of network failures.

November 25, 2025, Doha: The GSMA has announced its support for the AI ​​Telco Troubleshooting Challenge, which launched today in collaboration with ETSI, IEEE GenAINet, ITU and TM Forum. The new challenge calls for telecom operators, AI researchers and startups to come up with advanced large language models (LLM) that can perform root cause analysis (RCA) on telecom network failures. Network failures remain one of the most persistent and costly challenges facing the telecommunications industry, resulting in significant financial losses each year.

The new competition invites teams to submit AI models in three categories: Generalization to new faults will evaluate the best-performing LLMs for RCA; Small Models at the Edge will evaluate lightweight edge-deployable models; and Explainability and Reasoning will focus on AI systems that clearly explain their reasoning. Additional categories will include securing edge cloud deployments and enabling artificial intelligence services for application developers.

The goal of the challenge is to identify practical solutions that lead to more resilient, efficient and automated networks. The evaluation criteria for the challenge will evaluate participants on the accuracy, efficiency, reasoning ability, and security considerations of their models. In addition to leading community partners from ETSI, GSMA, IEEE GenAINet, ITU, and TM Forum, the AI ​​Telco Troubleshooting Challenge is supported by core partners Huawei, InterDigital, NextGCloud, RelationalAI, xFlowResearch, and technical advisors from AT&T.

“This challenge addresses some of the most pressing research questions in our industry, such as model generalization and the efficiency of edge-based AI,” said Dario Sabella, president of ETSI MEC (Multi-access Edge Computing). “By providing unparalleled access to specialized data and resources, we are accelerating the adoption of Telco AI. In particular, small and efficient language models, suitable for edge AI deployments, will reduce barriers to AI adoption, making advanced AI solutions more accessible and impactful across the industry.”

“Large language models have become critical in the quest for autonomous, resilient and adaptive networks,” said Professor Merouane Debbah, General Chair of the IEEE GenAINet ETI. “Through this challenge, we are addressing core research and engineering challenges, such as generalization to invisible network faults, interpretability, and efficient AI at the edge, which are vital to making AI-native telecom infrastructures a reality. IEEE GenAINet ETI is proud to support this initiative, which serves as a testbed for future-ready innovations across the global telecom ecosystem.”

“ITU's Global AI Challenges connect innovators with computing resources, data sets and expert mentors to nurture AI innovation ecosystems around the world,” said Seizo Onoe, Director of the ITU Telecommunication Standardization Bureau. “Our challenges drive business by helping innovations achieve meaningful impact by crowdsourcing new solutions and creating the conditions for them to scale.”

“The future of telecommunications depends on the autonomy of network resilience, moving from static infrastructure to AI-driven, context-aware and self-optimizing networks. TM Forum's AI-Native Blueprint provides the architectural foundation to make this a reality, and the AI ​​Telco Troubleshooting Challenge aligns perfectly to help the industry move from isolated pilots to production-grade resilient autonomy,” said Guy Lupo, AI and Data Mission Leader at TM Forum.

The initiative builds on recent advances in the application of AI to network operations, leveraging curated datasets such as TeleLogs and benchmarking frameworks developed by GSMA and its partners under the GSMA Open-Telco LLM Benchmarks community, which includes a league table highlighting how various LLMs perform in specific telecom use cases.

“Network failures cost operators millions a year and root cause analysis is a critical point for operators,” said Louis Powell, Director of AI Technologies at GSMA. “By leveraging AI models capable of reasoning and diagnosing invisible faults, the industry can dramatically improve reliability and reduce operating costs. Through this challenge, we aim to accelerate the development of LLMs that combine reasoning, efficiency and scalability.”

“We were encouraged by the benefits of this challenge after our team at AT&T fine-tuned a 4 billion-parameter small language model that outperformed all other models evaluated in the GSMA Open-Telco LLM (TeleLogs RCA Task) benchmarks, including cutting-edge models like GPT-5, Claude Sonnet 4.5, and Grok-4,” said Andy Markus, chief data officer at AT&T. “This challenge has the right combination of an important business problem and a technical opportunity, and we welcome industry collaboration to take it to the next level.”

AI Telco Problem Solving Challenge is open for submissions on the 28th^th November and closes on the 1st^street February 2026, with the winners announced in a session dedicated to the awards ceremony at MWC26 Barcelona.

Notes to editors

Replay of the global launch webinar of the AI ​​Telco Troubleshooting Challenge, ITU's AI for Good programme.