This post was originally published on Data Center Knowledge
Nvidia continued its partnership streak on Tuesday, announcing plans to team up with EPRI, Prologis, and InfraPartners to study smaller-scale data centers for distributed inference, which would be used for real-time processing across various sectors.
The partners will examine the deployment of so-called micro data centers – ranging from 5 to 20 MW – located at or near utility substations to enable quicker setup. The goal is to deliver inference capabilities near the point of data creation while optimizing underutilized infrastructure and easing strain on congested transmission systems, the companies said.
With five pilot sites planned across the US by year-end 2026, the collaboration aims to establish a replica model for rapid, scalable deployment.
“AI is transforming every industry, and the energy system will need to continue to evolve to meet increasing demand,” EPRI President and CEO Arshad Mansoor said in a release. “Using existing grid capacity to bring inference compute closer to where it’s needed – quickly and reliably – is a win for all.”
Inference, the process of using trained AI models to generate output, has different infrastructure requirements than training. Demand for AI inference is exploding as workloads shift
— Read the rest of this post, which was originally published on Data Center Knowledge.
