Supermicro is integrating NVIDIA Minerva, an AIOps platform launched at GTC Berlin that delivers end-to-end observability for AI factories, into the SuperCloud Software Suite — the management layer of Data Center Building Block Solutions®(DCBBS) — to help operators increase uptime and maximize useful output from every GPU hour.
AI factories generate enormous telemetry across GPUs, networking, storage, cooling, and workloads. A single degraded component can reduce goodput and token production. Traditional tools often leave teams with fragmented views. NVIDIA Minerva connects these signals into a dynamic knowledge graph, analyzes billions of data points per hour, detects anomalies, predicts failures, and correlates events into focused incidents with likely root causes. A native AI assistant and agentic AI integrations let operators investigate and remediate through natural language.
SuperCloud Composer® provides unified rack-scale and liquid-cooling management, power and leak detection, and performance telemetry for more than 20,000 hosts. The broader SuperCloud suite adds automation, multi-tenant control, and developer tools. Integrating NVIDIA Minerva extends this foundation with purpose-built AIOps focused on AI factory yield — combining infrastructure and workload insight so teams can move from reactive alerts to proactive maintenance.
Supermicro will offer NVIDIA Minerva as part of our AI factory operations portfolios, integrated with the SuperCloud Software Suite. Together with validated DCBBS infrastructure, this combination helps enterprises, neo-clouds, and hyperscalers achieve higher uptime and more useful work from every watt and GPU hour.
Learn more about NVIDIA Minerva and visit Supermicro at NVIDIA GTC Berlin.