en.Wedoany.com Reported - Circe Energy announced significant progress on its scalable behind-the-meter (BTM) AI and high-performance computing infrastructure platform. Through a strategic partnership with Cummins Inc., the company has ordered approximately 2 GW of natural gas power generation capacity, scheduled for delivery between 2026 and 2030, to support the phased development of its AI infrastructure campus in West Texas and future deployments across North America.
The flagship Circe project supported by the target generation capacity—the West Texas AI Infrastructure Campus—is located in the Permian Basin, spanning 1,950 acres, and is designed to support the deployment of gigawatt-scale AI and high-performance computing infrastructure. The campus integrates large-scale site control, competitively priced natural gas resources, advanced microgrid architecture, and HPC-ready power shell facilities into a single comprehensive development platform. Its proximity to a major North American energy production region helps achieve long-term scalability and competitive electricity costs.
As AI workloads continue to accelerate, power access has become a primary constraint for new data center development. Utility interconnection delays, supply chain shortages, transmission congestion, rising grid upgrade costs, and uncertainty in energization timelines are driving developers and hyperscalers to reassess traditional grid-dependent approaches.
Dagan Baroco, Chief Commercial Officer of Circe Energy, stated that AI infrastructure is fundamentally a power challenge. The market is recognizing that securing land or a utility queue position alone is insufficient; the key is the ability to energize when customers need it. Circe's platform is designed to address this issue.
The platform provides long-term electricity pricing stability through dedicated behind-the-meter natural gas generation, advanced microgrid architecture, long-term transportation and fuel supply arrangements, and power purchase agreements structured around predictable operating costs. By leveraging the abundant and competitively priced natural gas resources of the Permian Basin, customers can understand long-term energy economics and reduce exposure to volatile regional electricity markets.
Baroco added that customers are increasingly seeking solutions that prioritize speed, reliability, and certainty. The economics of waiting years for utility power, relying on expensive interim solutions, or bearing escalating grid upgrade costs are becoming increasingly difficult to justify. Behind-the-meter power allows investment to be directed straight into the data center itself, while reducing the costs, complexity, and uncertainty associated with utility interconnection and grid expansion.
By providing dedicated generation resources independent of local utility capacity constraints, Circe's model helps alleviate pressure on the existing grid while supporting regional economic growth through construction activity, long-term operational employment, and increased utilization of domestic energy resources.
Circe has completed significant development milestones in site control, permitting, engineering, fuel supply, generation procurement, and power infrastructure planning, supporting phased energization starting in 2027 and the long-term gigawatt-scale expansion of the West Texas AI Infrastructure Campus.
Circe Energy develops scalable behind-the-meter power and power shell infrastructure platforms for AI, high-performance computing, and mission-critical applications. By integrating assured generation supply, natural gas fuel, advanced microgrid architecture, land, and HPC-optimized building designs, Circe aims to provide predictable energization pathways, reduce grid risks, and accelerate deployment timelines for next-generation digital infrastructure.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
