NVIDIA is pushing telecommunications from simple automation toward true autonomy by introducing specialized AI tools designed to "think" like network engineers. Unlike traditional systems that follow rigid rules, these autonomous networks can understand an operator’s goals, weigh options, and decide on actions. At the heart of this shift is the new NVIDIA Nemotron-30B Large Telco Model (LTM), a massive AI brain fine-tuned on industry-specific data to handle complex tasks like fixing network faults and planning system changes.

To help companies adopt this technology, NVIDIA released "AI Blueprints," which act as architectural templates for building intelligent agents. These agents can manage specific tasks, such as reducing power consumption in 5G radio access networks (RAN) without hurting service quality. By using simulation tools to test ideas before applying them to the real world, these agents create a closed-loop system—a process where the AI constantly learns and refines its performance based on its own results.

Global operators are already putting these blueprints to work. In Africa, Cassava Technologies is deploying agents to optimize diverse mobile networks, while Telenor is enhancing maritime connectivity at sea. By open-sourcing these models through the GSMA’s Open Telco AI initiative, NVIDIA is providing the industry with the foundational tools needed to build self-managing, more resilient digital infrastructure.

NVIDIA is pushing telecommunications from simple automation toward true autonomy by introducing specialized AI tools designed to "think" like network engineers. Unlike traditional systems that follow rigid rules, these autonomous networks can understand an operator’s goals, weigh options, and decide on actions. At the heart of this shift is the new NVIDIA Nemotron-30B Large Telco Model (LTM), a massive AI brain fine-tuned on industry-specific data to handle complex tasks like fixing network faults and planning system changes.

To help companies adopt this technology, NVIDIA released "AI Blueprints," which act as architectural templates for building intelligent agents. These agents can manage specific tasks, such as reducing power consumption in 5G radio access networks (RAN) without hurting service quality. By using simulation tools to test ideas before applying them to the real world, these agents create a closed-loop system—a process where the AI constantly learns and refines its performance based on its own results.

Global operators are already putting these blueprints to work. In Africa, Cassava Technologies is deploying agents to optimize diverse mobile networks, while Telenor is enhancing maritime connectivity at sea. By open-sourcing these models through the GSMA’s Open Telco AI initiative, NVIDIA is providing the industry with the foundational tools needed to build self-managing, more resilient digital infrastructure.