Telecom operators adopt open-source AI and specialized tools

Telecommunications operators are increasingly adopting open-source artificial intelligence models to gain greater control over network automation, sensitive data, and customer workflows. According to Nvidia’s State of AI in Telecommunications 2026 survey, 89 per cent of operators view open-source software as critical to their AI strategies. Companies such as SoftBank Corp, AT&T, and Indosat Ooredoo Hutchison are utilizing a multi-model approach, reserving expensive proprietary systems for complex tasks while using lower-cost open foundation models for routine operations.

SoftBank has utilized Nvidia’s Nemotron family to develop its Large Telecom Model, integrating internal operational data to tailor the system to its specific network requirements. AT&T is also pursuing a multi-model strategy to match specific AI systems to various operational needs.

In a separate application of AI, AT&T has partnered with OpenAI to integrate AI tools, including chatbots and agents, into its in-house legal department. This initiative aims to streamline routine legal work and better utilize the company’s century of institutional legal knowledge. Through its internal law firm, LegalEdge, AT&T intends to reduce its reliance on external counsel, with plans to decrease its roster of national law firms from approximately 30 to between 18 and 20 over the next two years.