NVIDIA launches Minerva to monitor AI factory operations

NVIDIA has launched Minerva, an AIOps platform designed to detect emerging problems and improve productivity across AI factories. It combines telemetry from GPUs, networking, storage, schedulers and workloads, giving operations teams a shared view of infrastructure health and its impact on running jobs. The platform targets useful output measured through GPU hours, goodput, cluster uptime and token production.

Minerva builds a dynamic knowledge graph linking hardware components with the jobs running on them, helping operators trace performance issues to their likely source. Supervised and unsupervised machine learning identify abnormal behavior, predict emerging failures and correlate events using infrastructure topology. Related alerts can be consolidated into a focused incident rather than presented as separate symptoms.

A built-in AI assistant lets operators investigate cluster health and receive recommended fixes through natural-language conversations grounded in live telemetry. Teams can implement changes manually or connect the platform to automation through infrastructure application programming interfaces and external systems.

IREN is using Minerva to support proactive maintenance across its AI factories. Aolani, Bitdeer AI, IBM, Jio, Lambda, Quake AI and Tesla plan to deploy it. Dell Technologies, Lenovo and Supermicro will offer the platform within their AI factory operations portfolios.