NVIDIA® Minerva™ is an end-to-end, real-time AIOps platform purpose-built for AI factories and the teams who run them. Cloud service providers (CSPs) and enterprises use NVIDIA Minerva to increase AI factory yield through predictive maintenance, job failure root-cause analysis, noise reduction, and agentic operations capabilities. NVIDIA Minerva ingests full-stack telemetry across GPUs, NICs, DPUs, NVLink, and Spectrum-X Ethernet, applies purpose-built ML models and an AI Factory Assistant, and delivers measurable improvements in GPU hours, cluster uptime, token production, and MTTR.
Last updated: