● Product: Minerva, an AIOps platform for AI factories
● Inputs: GPU, networking, storage, scheduler and workload telemetry
● Stated scale: billions of data points per hour
● Planned deployments: IBM, Lambda, Tesla
● Pricing: not published
Nvidia launched Minerva, an AIOps platform that pulls telemetry from GPUs, networking, storage, schedulers and the AI workloads themselves, with the stated aim of flagging problems in AI factories earlier. Nvidia says it is designed to process billions of data points per hour. Planned deployments include IBM, Lambda and Tesla.
The strategic read is straightforward. Nvidia has sold an enormous installed base of accelerators into facilities that are difficult to operate, and operational software is both a second revenue line and a way to make that installed base harder to leave.
No pricing has been published. The named deployments are planned rather than live, which is worth holding in mind when the logos are used as evidence of adoption.
Sources
Tech Startups daily brief.