Nvidia Highlights Telecom Operators' Shift to Open-Source AI for Network Automation

Image: Illustrative image · Dicklyon · CC BY-SA 4.0 · Source

Nvidia reveals telecom operators like SoftBank, AT&T, and Indosat Ooredoo Hutchison increasingly use open-source AI models alongside proprietary ones to enhance network automation, data control, and cost efficiency.

Nvidia has emphasized its efforts to assist telecom operators in reducing their reliance on closed AI systems by promoting the adoption of open-source AI models. In a recent blog post, the company highlighted that operators such as SoftBank Corp, AT&T, and Indosat Ooredoo Hutchison are progressively applying open-source AI models in conjunction with proprietary systems.

This approach enables operators to maintain full visibility and control over sensitive network data while customizing AI models to meet their specific operational needs. It reflects a broader industry trend toward more transparent and adaptable AI deployments in telecommunications.

Supporting this shift, Nvidia shared findings from its "State of AI in Telecommunications 2026" survey, which found that 89% of operators consider open-source software critical to their AI strategy. Nvidia attributes this trend partly to independent benchmarking tests demonstrating how open models are narrowing performance gaps in reasoning, network programming, and autonomous classification tasks.

Operators are balancing workloads by reserving proprietary, often costlier models for complex tasks where closed platforms presently provide advantages. Meanwhile, routine operations are increasingly handled by more affordable open-source foundational models.

SoftBank, for example, utilizes Nvidia's Nemotron family as the foundation for its large-scale telecom AI models, trained on years of internal operational data. Rajeev Koodli, SoftBank's senior vice president and principal researcher, noted that open models allow them to leverage rapid global AI advances alongside their deep networking expertise.

Similarly, AT&T employs a multi-model strategy, aiming to assign workloads based on the best combination of performance, cost, and control. Andy Markus, AT&T's director of data and AI, highlighted this balanced approach as central to their AI deployment.

Indosat Ooredoo Hutchison has adapted open-source models to fit Indonesia's language and cultural context, demonstrating the flexibility such AI solutions offer across diverse markets.

These developments suggest a significant operator pivot towards open-source AI frameworks, seeking greater agility, transparency, and cost effectiveness in managing network automation and customer workflows.