A planned 10 GW AI data center campus tied to 9.2 GW of natural gas generation in Ohio is another clear sign that AI infrastructure is becoming an energy-first business. That is not a normal project announcement. It is a market signal. It is increasingly about who can secure power, deploy it fast, and structure around grid constraints. What also stands out here is where this is happening: a former uranium-enrichment site in Piketon. That may be a preview of what comes next — not just greenfield campuses, but repurposed legacy industrial sites where land, transmission potential, industrial zoning, and energy infrastructure can come together faster. For those of us in the data center and energy world, the bigger questions are: Does power availability now matter more than geography? Does behind-the-meter generation, turbines, microgrids, and BESS become standard for next-wave AI campuses? How many developers can actually execute on generation and electrical infrastructure fast enough to meet AI timelines? My takeaway: The winners in AI infrastructure may not just be the groups with the best sites. They may be the ones that can bring reliable megawatts to the table fastest. Does the future of AI infrastructure belong to traditional grid-connected campuses, or to energy-first developments built around dedicated power? #DataCenters #AIInfrastructure #EnergyInfrastructure #PowerGeneration #Microgrids #BESS #Hyperscale #DigitalInfrastructure