A shuttered paper mill in Jay, Maine is getting a second life as a data center. The 1.4 million-square-foot facility once employed 1,500 people before closing in 2020 after an explosion. Now it's being converted to house the servers and infrastructure that power AI systems.
This isn't an isolated case. Data centers are increasingly targeting rural America, drawn by available industrial sites, cheaper electricity, and local governments eager for economic development. These facilities need massive amounts of power and space, both of which are easier to secure outside major metro areas.
For AI users and builders, this shift matters because it's reshaping where compute happens. The infrastructure running your models, training your systems, and processing your queries is moving from traditional tech hubs to places like rural Maine. That geographic distribution could affect latency, costs, and availability.
Rural communities see data centers as economic lifelines, replacing lost manufacturing jobs with tech infrastructure roles. But the employment math is different. A paper mill employed 1,500 people. A data center might employ a few dozen once construction ends.
The trend also raises questions about power grid capacity and environmental impact. AI workloads are energy-intensive, and rural grids weren't built for this kind of demand. As more training runs and inference requests flow through these facilities, the strain on local infrastructure will grow.
For anyone building with AI at scale, watching where data centers locate tells you something about the future cost and availability of compute. Rural expansion suggests the industry is hitting constraints in traditional markets and needs to spread out to keep growing.
