The Electric Power Research Institute (EPRI) has embarked on a collaboration with Prologis, NVIDIA, and InfraPartners to examine the feasibility of micro data centres. These scaled-down facilities, ranging from 5 to 20 megawatts, are set to enhance real-time data processing through distributed inference across diverse sectors including logistics, healthcare, finance, and public services.
Announced at DTECH, the initiative aims at computational innovation and utility of available infrastructure. By stationing these data centres near utility substations with unused grid capacity, the collaboration intends to provide rapid responses where data is generated, reducing the strain on existing transmission systems.
The project targets the development of at least five pilot sites across the US by 2026. This model aims for scalability and swift deployment, enabling sectors to adapt to AI-driven transformations.
AI's transformative potential extends to redefining how industries utilise energy, this collaboration seeks to contextualise and apply inference compute close to data sources.
EPRI plays a role by identifying locations and gathering insights from active projects. As Prologis explores viable land and building solutions for these data nodes, they aim to drive commercialisation through orderly planning and energy acumen.
NVIDIA's contribution comes in the form of GPU-accelerated platforms for distributed inference, while InfraPartners supplies AI data centres with high-density power solutions.
Each collaborator seeks to streamline the approach to meet the demands of the burgeoning digital economy.
By integrating compute capacity proximal to existing power capabilities, this initiative enhances grid reliability while leveraging established infrastructure for digital advancement. The positioning of micro data centres helps alleviate congestion within transmission systems and supports the integration of renewable energy sources.
The collective efforts of EPRI, Prologis, NVIDIA, and InfraPartners aim to respond proactively to increasing industry demands for smarter infrastructure solutions.
With AI driving demand for distributed computing, the partnership focuses on deploying infrastructure that leverages existing resources. The collaboration provides a model for scaling AI workloads and supporting real-time applications across multiple sectors.
