Meta Pushes Ahead With AI Infrastructure Expansion
Meta is seeking $3 billion in financing to support the construction of an artificial intelligence data center called Prometheus in New Albany, Ohio. The project is designed with a planned capacity of 1 gigawatt, and the financing package covers both the data center itself and the associated power generation assets. The reported pricing is set at SOFR + 2.5%.
The move highlights Meta’s continued investment in large-scale AI infrastructure as global demand for computing power keeps rising. As AI training and inference workloads become more intensive, major technology companies are increasingly racing to secure not only data center capacity, but also reliable energy supply to support those operations.
Dedicated Gas Power at the Core of the Project
A defining feature of Prometheus is its independent gas-powered electricity supply. This setup would allow the facility to operate like an islanded microgrid, reducing reliance on the external grid and helping maintain service even during broader grid outages.
That approach reflects a growing industry focus on power resilience. AI-focused data centers generally require far more energy and greater operational stability than conventional facilities. By combining data center development with on-site energy infrastructure in the same financing plan, Meta appears to be exploring a more integrated model for future AI buildouts.
EdgeConneX Involved in Development
Infrastructure firm EdgeConneX is participating in the construction of the project. Known for its experience in digital infrastructure, the company’s involvement could support execution of a development of this scale.
More broadly, the Prometheus initiative is aimed at meeting the expanding global need for AI compute while testing a more energy-autonomous model for data centers. If successful, the project could offer a new benchmark for future facilities, particularly in areas such as uptime, grid independence, and the integration of dedicated power systems with next-generation AI infrastructure.