Telecom Operators and AI: The Role of Open Models

More and more telecom operators are basing their AI strategies on open models, with benefits that go beyond just cost savings. Open models provide telecom companies the opportunity to trust, manage, and adapt AI for their most critical workloads, from autonomous networks to customer service. According to the latest State of AI in Telecommunications report from NVIDIA, 89% of respondents consider open-source models and software essential for their AI strategy.

The strategic value of open models for telecom operators is fivefold. They expand access to advanced intelligence at lower costs, and support telco-specific adaptations with open weights and training recipes. Additionally, open models increase trust in AI by offering better insight and control over model artifacts and behavior, which is essential for governance and regulatory compliance. Flexibility in deployment allows models to be securely rolled out across various infrastructures and applications, and they form the foundation for locally tailored AI services for government and enterprise clients.

The NVIDIA Nemotron family of open models offers advanced reasoning capabilities and speech functionalities, with open weights and training data. Operators such as SoftBank Corp. use these models as a basis for developing telecom-specific AI capabilities. NVIDIA collaborates with partners like AdaptKey to convert open models into usable building blocks for telecom AI, including the Nemotron 3 Large Telco Model, which is fine-tuned for telecom-specific tasks.

To bring the benefits of open models into production, telecom operators need robust pipelines to protect and prepare their data for model tuning. NVIDIA provides an end-to-end platform that facilitates building autonomous telecom operations, supported by their AI Enterprise software and Agent Toolkit. This enables telecom operators to successfully translate the value of open models into production-ready AI workflows and adapt to local languages and industries. It is a clear pathway for these companies to realize a strong AI infrastructure that meets national strategies.

Read the full article from Nvidia.