Trump Admin Unveils 10GW AI Data Center & Gas Power Hub at Ohio Uranium Site

The announcement of a 10-gigawatt (GW) data center and power complex in southern Ohio marks a watershed moment in the intersection of national security, energy policy, and the global AI arms race. By repurposing a former uranium enrichment site—a relic of the Cold War—the Trump administration is signaling a "bricks and mortar" approach to AI dominance. The scale of 10 GW is unprecedented; for context, a typical large data center consumes about 100-500 MW. This project is equivalent to the output of ten large nuclear reactors, highlighting the staggering energy appetite of next-generation generative AI models and the administration's commitment to securing the physical infrastructure required to lead in the digital age.

The partnership between the U.S. Department of Energy (DOE), SoftBank, and AEP Ohio illustrates a fundamental shift in how hyperscale infrastructure is being permitted and powered. While tech giants like Microsoft and Google have historically leaned toward renewable energy credits to offset their carbon footprints, the sheer scale of this Ohio project necessitates a "behind-the-meter" or dedicated gas-fired power solution. This aligns with the administration's broader push to deregulate fossil fuel production and prioritize grid reliability over aggressive decarbonization targets. SoftBank’s involvement, led by Masayoshi Son, underscores the company's pivot from a venture capital powerhouse to an AI infrastructure giant, seeking to secure the physical compute capacity that will define the next decade of technological competition.

“

Department of Energy (DOE), SoftBank, and AEP Ohio illustrates a fundamental shift in how hyperscale infrastructure is being permitted and powered.

The choice of a former uranium site—specifically the Portsmouth Gaseous Diffusion Plant area—is geographically and logistically strategic. These sites often possess existing high-voltage transmission infrastructure and access to significant water resources, which are the two primary bottlenecks for data center development today. By utilizing a brownfield site that has been undergoing federal cleanup for decades, the project effectively bypasses many of the greenfield development hurdles that plague similar large-scale projects. However, the reliance on natural gas plants to power the site will likely draw scrutiny from environmental groups and potentially complicate long-term sustainability goals, even as it provides the reliable baseload power that AI training clusters require.