The collaboration between EPRI with Nvidia, Prologis and InfraPartners will prioritise smaller-scale data centres designed for distributed inference, a form of real-time data processing used across sectors including logistics, health care, finance and public services.

Assessing the deployment of micro data centres, ranging from 5-20MW, the companies will explore how these smaller distributed sites could meet rising computing needs without straining the grid. They will be looking at these sites at or near utility substations with available grid capacity that can be set up quickly.

The goal is to bring inference capabilities closer to where data is generated and consumed, while making better use of infrastructure that is under-utilised and also reducing pressure on congested transmission systems.

Hoping to have at least five pilot sites in development across the US by the end of 2026, the collaboration aims to provide a replicable model for rapid and scalable deployment.

“AI is transforming every industry, and the energy system will need to continue to evolve to meet increasing demand,” said EPRI president and CEO Arshad Mansoor. “This collaboration with Prologis, Nvidia, InfraPartners and the utility community highlights the type of innovative actions required to meet the moment.”

Mansoor added: “Using existing grid capacity to bring inference compute closer to where it’s needed – quickly and reliably – is a win for all.”

Each organisation will play a key role in supporting this project and the development of sites that relieve pressure on the grid. This includes:

EPRI: Providing independent, unbiased technical expertise and research-based validation.

Prologis: Identifying and evaluating suitable land and buildings near substations for rapid deployment, while coordinating development, planning, permitting and delivery as demonstration sites advance.

Nvidia: Delivering optimised, GPU-accelerated computing platform for distributed inference workloads, in addition to offering technical guidance on distributed inference architectures and facilitates connections to potential compute offtakers.

InfraPartners: The AI data centre builder, providing facilities designed for high-density power and cooling and built using advanced offsite manufacturing.

Participating utilities: Assessing available distribution capacity, informing sitting and interconnection pathways and ensuring alignment with operational requirements.

As AI applications scale across industries, demand for AI inference continues to surge. The companies argued that meeting this demand not only requires more computing, but AI infrastructure deployed closer to end-users will relieve pressure on congested transmission systems.

“As energy demand grows, we need infrastructure solutions that support grid reliability and make better use of what’s already built,” said Parag Soni, senior vice president and global head of Utility Strategy and Engagement at Prologis. “This collaboration is about using our development and energy expertise to help deliver smarter, more flexible infrastructure right where it’s needed.”

This approach supports grid reliability. By co-locating computing capacity with substations that have existing but underused distribution headroom, utilities may reduce transmission congestion, improve system flexibility and help integrate renewable energy.

“AI is driving a new industrial revolution that demands a fundamental rethinking of data centre infrastructure,” said Marc Spieler, senior managing director for the Global Energy Industry at Nvidia. “By deploying accelerated computing resources directly adjacent to available grid capacity, we can unlock stranded power to scale AI inference efficiently.

“This distributed approach, powered by Nvidia accelerated computing, maximises existing energy assets, helping to deliver the intelligence required to transform every industry.”

Harqs Singh, chief technology officer at InfraPartners, added: “AI is becoming the real-time engine of growth for the modern economy and it demands a new kind of digital infrastructure.

“AI data centre solutions with EPRI’s technical leadership, Nvidia’s platforms and Prologis’ national footprint, we’re enabling rapid deployment of AI nodes where they’re needed most. Together, we’re building the foundation for the next decade of intelligent infrastructure.”

Related stories

Nvidia, Microsoft and Amazon in talks on up to $6bn OpenAI investment

Data centre of the month: Tecto Data Centers’ TPOA1

Microsoft, Nvidia and tech giants pledge over $40bn to UK AI infrastructure and data centres