NVIDIA AI Enterprise is used by Amdocs, BubbleRAN, ServiceNow, SoftBank, and Tech Mahindra to create new network AI agents and massive telecom models.

Millions of user connections may be supported daily by global telecommunications networks, which on average produce over 3,800 terabytes of data each minute. Network traffic data, performance metrics, configuration, and topology are all part of the vast, constant stream of unstructured, complicated data produced by base stations, routers, switches, and data centers. It should come as no surprise that enormous, real-time tasks requiring such data have frequently proven too difficult for typical automation technologies to handle.

At the GTC global AI conference, NVIDIA revealed that its partners are employing NVIDIA NIM and NeMo microservices within the NVIDIA AI Enterprise software platform to create new large telco models (LTMs) and AI agents specifically designed for the telecom sector in order to assist address this difficulty. The next generation of AI in network operations is made possible by these LTMs and AI agents.

The creation of network AI agents, which automate intricate decision-making processes, increase operational effectiveness, increase worker productivity, and improve network performance, relies heavily on LTMs, or customized, multimodal large language models (LLMs) trained especially on telco network data.

While Amdocs, BubbleRAN, and ServiceNow are utilizing NVIDIA AI Enterprise to enhance their network operations and optimization, SoftBank and Tech Mahindra have developed new LTMs and AI agents. It's crucial work at a time when 40% of participants in a recent telecom poll conducted by NVIDIA stated they are using AI in their network operations and planning.