NVIDIA unveils open model strategy for telecom operators and Nemotron 3 LTM

Why operators are turning to open models

According to NVIDIA's new State of AI in Telecommunications report, 89% of telecom operators say open source models and software are important to their company's AI strategy. NVIDIA argues the reason goes beyond cost: open models give operators more control, let them adapt AI with their own data, and make it usable in critical workloads.

The company lists five strategic benefits for operators: access to frontier-level intelligence at lower cost, telecom-specific customization, more transparent and governable AI, flexible deployment across public cloud, private infrastructure and edge environments, and the ability to offer locally adapted AI services to enterprises and governments.

Nemotron 3 Large Telco Model

NVIDIA announced the 30-billion-parameter Nemotron 3 Large Telco Model (LTM), a telecom version of its open-weight Nemotron family. Fine-tuned by AdaptKey on open source telecom datasets, the model aims to improve accuracy for telecom tasks such as network configuration and customer incident triage.

To help operators adapt the model with their own network, customer and process data, NVIDIA also released an end-to-end fine-tuning recipe using NVIDIA NeMo open libraries.

SoftBank and AT&T examples

SoftBank Corp. says it uses open foundations, including NVIDIA Nemotron models, to develop its "SoftBank Large Telecom Model." AT&T, meanwhile, stresses that model choice should be made operational in line with business priorities.