NVIDIA: Open AI Models & Blueprints Power Autonomous Telecom Networks

Barcelona – NVIDIA unveiled a new open-source large telco model (LTM) and a suite of AI blueprints designed to accelerate the development of autonomous networks, ahead of this week’s Mobile World Congress. The move signals a growing industry focus on leveraging artificial intelligence to automate and optimize telecommunications operations.

According to NVIDIA’s recent State of AI in Telecommunications report, network automation is the top AI investment priority for telecom operators. However, the company emphasizes a distinction between automation – executing pre-defined tasks – and true autonomy, which requires networks to understand operator intent, weigh options, and independently decide on actions.

The newly released NVIDIA Nemotron-based LTM, a 30-billion-parameter model developed in collaboration with AdaptKey AI, is intended to provide the reasoning capabilities necessary for autonomous networks. It’s trained on open telecom datasets, including industry standards and synthetic logs, to understand the specific language and workflows of the telecommunications industry. The model is designed to be deployed on-premises, giving operators control over their data and security.

NVIDIA is releasing the LTM, an implementation guide, and agentic AI blueprints as open resources through GSMA, an organization representing the interests of mobile network operators worldwide, as part of the organization’s new Open Telco AI initiative launching Tuesday.

Alongside the LTM, NVIDIA and Tech Mahindra have published a guide detailing how to fine-tune reasoning models and build agents capable of executing network operations center (NOC) workflows. The guide focuses on teaching AI to mimic the reasoning process of experienced network engineers, translating expert resolutions into structured, step-by-step procedures for the model to learn from.

The company also introduced a new Blueprint for intent-driven RAN energy efficiency, integrating network test and measurement technology from VIAVI to generate synthetic network data. This allows operators to simulate and validate energy-saving policies without disrupting live networks.

Several telecommunications companies are already adopting NVIDIA’s Blueprint for telco network configuration. Cassava Technologies is utilizing the blueprint to build Cassava Autonomous Network, aiming to optimize its mobile network environment across Africa. NTT DATA is implementing the blueprint to intelligently manage traffic surges and improve network resilience for a tier 1 operator in Japan.

Further enhancing these capabilities, NVIDIA and BubbleRAN are collaborating to integrate NVIDIA’s NeMo Agent Toolkit (NAT) and BubbleRAN’s Agentic Toolkit (BAT) into the Blueprint for telco network configuration. This integration aims to provide more flexible orchestration of multi-agent workflows across the radio access network (RAN). Telenor Group will be the first to adopt this enhanced blueprint to improve its 5G network for Telenor Maritime.

NVIDIA also announced a commitment, alongside Booz Allen, BT Group, Cisco, Deutsche Telekom, Ericsson, MITRE, Nokia, OCUDU Ecosystem Foundation, ODC, SK Telecom, SoftBank Corp., and T-Mobile, to build the next generation of wireless networks – 6G – on AI-native, open, and secure platforms. The initiative aims to ensure that 6G infrastructure is intelligent, resilient, and fosters innovation.