Mircea Anghel, director of service operations at O2 Telefonica, said the model is intended to bring more automation into network operations
In sum – what to know:
O2 Telefonica develops Large Telco Model – A new generative AI model aims to improve fault detection, technician coordination, and overall service quality in its German network.
Partnership with Tech Mahindra and Nvidia – The system is based on Nvidia AI Enterprise software and agent-based AI, adapted specifically for telecommunications operations.
Toward autonomous network operations – Executives from O2, Tech Mahindra and Nvidia highlight the model’s role in advancing automation, data-driven processes and more reliable customer experiences.
German carrier O2 Telefonica is working with Tech Mahindra and Nvidia with the goal of developing a generative AI model designed for telecom networks, aimed at improving operations efficiency and customer service reliability, the former said in a release.
The project, called the “Large Telco Model,” applies generative AI and automation to network management. The system is intended to help technicians identify disruptions faster, coordinate field operations more effectively, and enhance overall service quality.
Mircea Anghel, director of service operations at O2 Telefonica, said the model is intended to bring more automation into network operations and improve decision-making with data-driven processes. He added that the approach aims to increase reliability and efficiency while giving operations teams more autonomy.
Tech Mahindra’s chief transformation officer, Amol Phadke, said: “Autonomous operations are becoming increasingly important for telecom operators. We’re excited to be working with O2 Telefonica and their ‘Operations of the Future’ initiative to bring these agent-based AI capabilities to life. By developing O2 Telefonica’s Large Telco Model, we ‘re leveraging the power of our AI platform and advanced agent-based capabilities to deliver real-world solutions that create measurable value.
The model, based on Nvidia AI Enterprise software, was first presented at Nvidia’s GTC conference in March 2025. It uses both structured data, such as network alarms and meter readings, and unstructured data, such as logs, images, and text. By combining O2 Telefónica’s network data with advanced AI models, the project aims to move towards more autonomous telecom systems, the telco said.
O2 Telefonica added the collaboration reflects its strategy to advance data-driven operations and position itself as an early adopter of AI in Germany’s telecom sector.
In March, O2 Telefónica had selected Amazon Web Services (AWS) as its cloud provider for network workloads.
The European operator noted that this multi-year agreement, signed during Mobile World Congress (MWC) 2025, builds on the previous deployment of O2 Telefónica’s 5G SA Core on AWS in July 2024. As part of this expansion, O2 Telefónica will extend its 5G core implementation on AWS, deploying the 5G core data plane within its data centers using AWS Outposts racks to support network-intensive workloads. Additionally, the company will begin deploying IMS voice services on AWS Graviton processors this year.
