Abandoned U.S. factories are being converted into AI data centers as companies seek faster buildout timelines

Across parts of the United States where manufacturing jobs have disappeared over decades, a new wave of infrastructure investment is targeting the physical shells those industries left behind. WhiteFiber, an AI infrastructure company, is among the firms converting former factory buildings into data centers to support the growing demand for AI computing power.

Sam Tabar, CEO of WhiteFiber, told CNN that sites like the shuttered textile plant in Madison, North Carolina represent what he described as a "secret weapon" in the race to build AI capacity. His argument centers on practical advantages: abandoned industrial facilities often already have heavy electrical infrastructure, large floor plates, and structural foundations capable of supporting dense computing equipment — features that can take years to build from scratch at a greenfield site.

The Madison plant had been vacant since it shed 250 jobs, a pattern common across former textile and manufacturing corridors in the American South and Midwest. Tabar said converting such sites can reduce the time to operational readiness from years to months compared with new construction.