French energy and water utility Veolia has been selected to operate and maintain a 350MW microgrid that will power an undisclosed AI data center campus in Ohio, marking one of the largest behind-the-meter energy projects tied to the AI infrastructure buildout in the United States.
The energy project is expected to provide 100 percent of the data center’s power demand. The microgrid will be anchored by gas engines and linear generators, with the capacity supplemented by a large-scale battery energy storage system (BESS) rated at 430MWh. This configuration allows the facility to operate fully behind the meter, with no reliance on the grid, a critical feature as data center developers face prolonged interconnection queues and grid capacity constraints across major U.S. markets.
“Our work on this project shows how Veolia can help developers accelerate speed-to-power while maintaining the reliability required for mission-critical AI operations, all while protecting the community,” said Karin Hamel, CEO of Veolia Energy, North America and chief growth officer for Veolia in North America. “As developers deploy microgrids and other advanced energy systems to meet the growing demand for AI infrastructure, they need trusted partners with the expertise to operate these assets safely, reliably and at scale.”
Veolia will oversee operations integration, commissioning support and long-term plant operations under a performance-based model, which the company says is designed to maximize uptime, asset performance and operational accountability. According to Veolia, the company was brought onto the project before commercial operations to support startup readiness, operating procedure development and coordination among multiple contractors and original equipment manufacturers working under compressed construction timelines.
The contract builds on Veolia’s Data Center Resource 360 offer, launched earlier this year. The platform is designed to support the development of data centers into carbon-neutral, water-positive, circular hubs, with claims of up to 20 percent more energy reuse, a 75 percent smaller water footprint, and 95 percent waste recycling and reuse.
The project underscores a broader industry shift toward self-generation and microgrid solutions as hyperscale and AI-focused developers seek to bypass grid bottlenecks and secure reliable power on accelerated schedules. As the demand for AI compute continues to surge, the ability to deliver dedicated, dispatchable power behind the meter is becoming a decisive factor in site selection and project viability, particularly in regions like Ohio where grid capacity is increasingly contested.