NVIDIA Minerva tracks AI factory failures with a knowledge graph
NVIDIA’s new AIOps platform Minerva combines GPU, networking, storage and workload telemetry in a dynamic knowledge graph to detect anomalies and potential failures.
AT A GLANCE
Key points
- A dynamic knowledge graph maps relationships across GPUs, networking, storage, schedulers and workloads.
- The platform analyzes billions of data points per hour for anomaly detection and failure prediction.
- A native AI assistant provides explanations and remediation suggestions grounded in live telemetry.
- IREN is already using it, while IBM, Tesla and others plan deployments.
What is changing?
NVIDIA has introduced Minerva, an AIOps platform designed to simplify the operation of large-scale AI factories. It brings together infrastructure and workload telemetry spanning GPUs, network switches, storage and schedulers. The data is turned into a dynamic knowledge graph that maps relationships between components and jobs.
Minerva analyzes billions of data points per hour using supervised and unsupervised machine learning. It learns normal behavior patterns to identify new anomalies, predict failures and correlate related events across the physical topology. The goal is not merely to generate alerts, but to narrow down root causes faster.
A native AI assistant answers natural-language questions grounded in live telemetry and recommends remediation steps. Operators can apply those steps with human approval or connect them to automation through infrastructure APIs.
IREN is already using the platform, while Aolani, Bitdeer AI, IBM, Jio, Lambda, Quake AI and Tesla plan deployments. Minerva will also be available with systems from Dell Technologies, Lenovo and Supermicro.
Explore the original sources
Source date: 22 October 2026.
Published by AI Radar: 22 Oct 2026, 19:06 (Istanbul, UTC+3). Updated: 22 Oct 2026, 20:06.
FOLLOW THE TOPIC
Topics and tags
Share your discovery.
Send this story to a friend.
Reader comments
What do you think about this development? Share your experience or question. Comments are published after editorial approval.
Your name/nickname and comment become public if approved. Read: Privacy notice.
Loading comments…