AI agents trained on sector-specific documents are expected to help utilities assess energy needs quickly, potentially reducing interconnection study timelines by over 80%.

Led by the Electric Power Research Institute (EPRI), the consortium includes various energy and technology companies and aims to develop open models using industry-specific data.

This consortium, formed by Nvidia, Microsoft, AWS, and Oracle, seeks to enhance electric power generation and distribution by promoting AI adoption.

The Executive Advisory Committee of the consortium comprises leaders from over 20 energy companies and major tech firms, with plans to expand its global membership.

On March 20, 2025, the Open Power AI Consortium was launched at the NVIDIA GTC conference in San Jose, California, with the aim of transforming the energy sector through AI.

The initiative aims to enhance data organization and usability for AI applications in the energy sector, according to EPRI's Jeremy Renshaw.

Over two dozen U.S. regional power companies, including Con Edison and Duke Energy, have already joined the consortium.

EPRI President Arshad Mansoor highlighted that the consortium will focus on developing AI models, datasets, and applications to improve grid reliability and energy management.

Collaboration is a key focus, as the consortium seeks to engage utilities, startups, academics, and national labs to address challenges in the power sector.

PPL Corporation CEO Vincent Sorgi, who also serves as EPRI's board chair, expressed a commitment to leveraging AI for operational excellence in the energy sector.

The consortium has developed the first version of an open AI model for electric systems using NVIDIA H100 GPUs, which will soon be available for early access as an NVIDIA NIM microservice.

Additionally, the consortium plans to create a standardized framework to evaluate the performance and reliability of AI technologies within the energy sector.