Tech Mahindra Announces LLM For Telecom Operators

IT services giant Tech Mahindra has announced a new multimodal large language model (LLM) for telecom operators. 

Developed on top of NVIDIA AI Enterprise software and AWS cloud infrastructure, the LLM will be based on Meta’s Llama 3.1 8b instruct model. 

In a statement, Tech Mahindra said that the new model will pave the way for the transformation of traditional telecom networks into fully autonomous networks (L4 and above). 

The company also noted that the LLM will be built by utilising the latest generative AI (GenAI) and agentic AI frameworks, and will be trained on large network datasets. The new model will be designed to manage structured, unstructured and other relevant network data to enable “proactive issue resolution and enhanced service quality”.

“… Our collaboration with NVIDIA and AWS is pioneering a multi-modal network operations large model designed to enhance security, automate network management, and improve operational efficiency. Through this work, we will empower telcos to reduce operational costs and pave the way for a more agile and resilient network environment,” Tech Mahindra’s CTO for telecom business Manish Mangal said.

During the initial phase of the development, the IT services major said that the LLM will improve operational efficiency for telcos by offering an AI-powered network observability solution and an AI-driven auto-resolution system that will autonomously detect and resolve network anomalies without “human intervention”.

In addition, Tech Mahindra added that the solution architecture will also integrate “AI-driven intelligence” into network operations across three components, including efficient data ingestion from the network, data curation and model customisation to enhance AI training, and automated quick resolution and restoration of services.

“… Large telco models like Tech Mahindra’s new multi-modal network operations large model — based on NVIDIA AI Enterprise — offer the foundation for creating multiple AI agents that will help enable fully autonomous networks,” Nvidia’s vice president of telco business development Chris Penrose said.

Tech Mahindra also noted that the latest offering is in line with its larger strategy of leveraging multimodal LLMs for building other business use cases as well. The announcement came on the sidelines of the ongoing Mobile World Congress 2025 in Spain’s Barcelona. 

A day earlier, Jio Platforms said that it has partnered with AMD, Nokia and Cisco to build an open-source AI platform for telcos, which will integrate AI and automation into every layer of telecom network operations. 

Meanwhile, SaaS major Zoho is also developing its own Indic language-focussed LLM.