NVIDIA Minerva
NVIDIA launched Minerva, an AIOps platform that joins infrastructure and workload telemetry to find problems earlier and keep AI factories productive.
Minerva builds a dynamic knowledge graph of GPUs, SuperNICs, switches, cables and the jobs running on them, so a slow training job can be traced to the component causing it. It uses supervised and unsupervised machine learning on billions of data points per hour to detect anomalies, predict failures and group event storms into single incidents with a likely root cause. It includes a native AI assistant and connects to external agentic tools and automation through infrastructure APIs.
- Date
- Thursday 22 October 2026
- Lab
- NVIDIA
- Kind
- product
- Access
- closed API
Figures
| Measure | Value | Measured by |
|---|---|---|
| Data points analyzed | billions per hour stated for the platform's analysis of telemetry | company |
IREN is using it. Aolani, Bitdeer AI, IBM, Jio, Lambda, Quake AI and Tesla plan to deploy it. Dell, Lenovo and Supermicro will offer it. Access type is not stated on the page, so the access value is a best fit.
Sources
This record was checked against its sources on 22 October 2026. How we check