NVIDIA Launches Minerva AI Factory Platform
The AIOps platform unifies telemetry across GPUs, networking, storage, and jobs to manage large-scale AI infrastructure.
Updated on Nov. 5, 2026 in Artificial Intelligence
NVIDIA has launched Minerva, an AIOps platform designed to manage AI factory operations. IREN already uses the platform, and Aolani, Bitdeer AI, IBM, Jio, Lambda, Quake AI, and Tesla plan to deploy it.
Why it matters
AI factories can involve hundreds of thousands of GPUs linked across high-speed networks, making unified operational views difficult to maintain. Expensive computing assets must keep running at useful output for as long as possible, which makes faster fault detection and repair central to operator economics.
Minerva combines telemetry from GPUs, networking, storage, schedulers, and jobs, and uses supervised and unsupervised machine learning to process billions of data points an hour. The extraction does not include independent validation of these throughput claims or measured results from deployments.
The players
NVIDIA
NVIDIA is a US semiconductor company best known for its graphics processing units, which have become central to training and running large-scale artificial intelligence systems.
Dell Technologies
Dell Technologies is a US computing hardware and infrastructure company that sells servers and storage systems used in enterprise and data center deployments.
Lenovo
Lenovo is a China-headquartered multinational technology company that manufactures personal computers, servers, and data center hardware.
Supermicro
Supermicro, formally Super Micro Computer, is a US manufacturer of server and storage hardware widely used in high-performance computing and AI infrastructure.
IBM
IBM is a US multinational technology company with long-standing enterprise software, consulting, and hybrid cloud businesses.
The details
Minerva builds a dynamic knowledge graph mapping GPUs, SuperNICs, switches, and cables to the jobs using them, then generates alerts, groups related signals into incidents with a likely root cause, and suggests remediation. Dell Technologies, Lenovo, and Supermicro will offer the platform within their AI factory operations portfolios, and it includes a native AI assistant that can connect to external agentic AI tools.
Timeline
- 2026-11-05: NVIDIA launched Minerva, as reported by datacenter.news.
The Tech Race
Minerva extends NVIDIA's position beyond the accelerator chips that built its lead into the operational software that keeps large GPU clusters running. Rival hardware makers and cloud providers are now competing on how efficiently their AI infrastructure can be operated, not only on raw compute.
For organizations running AI workloads, Minerva offers a single view of GPUs, networking, storage, and jobs, with alerts and suggested fixes that could shorten troubleshooting. Availability depends on deployment through Dell Technologies, Lenovo, Supermicro, or the other named partners rather than a standalone purchase.
The takeaway
Operators running large AI clusters should treat uptime and repair speed as core performance metrics alongside raw compute. Evaluating how any AIOps tool maps infrastructure to active jobs is a practical starting point.
Further reading
Explore more coverage in our Artificial Intelligence section for context on how AI infrastructure is evolving.
Source note: This article includes information reported by DataCenterNews US.
