NVIDIA Launches Minerva AI Factory Operations Platform, Capable of Analyzing Billions of Data Points Per Hour

October 22 — NVIDIA has launched Minerva, an AI operations platform designed to detect failures in advance, increase GPU uptime and boost AI factory output efficiency by consolidating telemetry data from infrastructure and AI workloads. The platform can analyze billions of data points per hour, identify anomalies, predict potential failures and consolidate related alerts into actionable incident items. Minerva also features a native AI assistant that lets users query cluster health in natural language and receive remediation suggestions.

NVIDIA said IREN is already using the platform, with Aolani, Bitdeer AI, IBM, Jio, Lambda, Quake AI and Tesla planning deployments; Dell, Lenovo and Supermicro will also incorporate Minerva into their respective AI factory operations portfolios.