The Ohio AI Data Center Boom: What a 10-Gigawatt Campus Means for Power, Cooling, and Controls

Nvidia is in talks to provide roughly $250 billion in financing guarantees for OpenAI's lease of a 10-gigawatt AI data center campus in southern Ohio. The Wall Street Journal broke the story on July 26, and Reuters confirmed it the same day. One widely shared summary on X captured the scale: more than $500 billion in total cost including the chips, a first phase of around 800 megawatts targeted for 2028, and separate talks about Nvidia financing up to $350 billion in OpenAI chip purchases.

The campus is being developed by SB Energy, a SoftBank subsidiary, on the former Portsmouth Gaseous Diffusion Plant site near Piketon in Pike County. The Department of Energy announced the partnership with SoftBank and AEP Ohio in March. The site enriched uranium through the Cold War, and it offers what almost no greenfield site can: federal land, existing transmission, and room for gigawatt-scale generation. None of this is signed yet, and other tenants are reportedly in the conversation. But the campus is moving, and it is the clearest signal yet of where Ohio sits in the AI buildout.

Ohio did not stumble into this. Amazon Web Services, Google, and Microsoft have all made major data center investments in the state, and central Ohio recently added more hyperscale capacity than any other US region over an 18-month stretch, something we covered in our piece on Ohio's water-compliance battleground. The advantages keep stacking up:

• Reliable, scalable power. Transmission capacity and the ability to add generation matter most at AI scale. The Piketon plan includes new gas-fired plants for exactly that reason.
• Location and fiber. Ohio sits within a day's reach of most of the eastern US, on long-established fiber routes.
• Cost structure. Land, power, and labor all price below coastal markets.
• Industrial land. Legacy sites like Piketon carry infrastructure a developer could never permit from scratch on this timeline.

Traditional data centers were already complex facilities. AI-focused ones add several layers of demand on top. Power density per rack climbs dramatically. Air cooling gives way to liquid and hybrid systems. Uptime, redundancy, and failover stop being IT concerns and become facility engineering problems.

Strip away the servers and look at what actually has to be monitored and controlled, and the list reads like the equipment schedule of an industrial plant:

• Power distribution and backup generation. Switchgear, transfer schemes, generator sequencing, and load management, all instrumented and automated.
• Cooling infrastructure. Chilled water plants and liquid cooling loops are process systems: pumps, valves, flow, pressure, and temperature instrumentation, with control logic holding tight setpoints around the clock.
• Environmental monitoring. Temperature, humidity, and airflow, measured continuously and alarmed intelligently.
• Building management and energy optimization. The BMS ties it together, and at AI scale, efficiency percentage points are worth real money.

And underneath all of it sits water. Cooling at this scale is a water story, which is why data center water use keeps drawing regulatory attention. We have covered that in depth, from the emerging 2026 water standards to the compliance pressure building in Ohio specifically.

A 10-gigawatt AI campus is not an IT project with a big electric bill. It is an industrial facility, and it will succeed or struggle on industrial fundamentals: instrumentation, controls, and monitoring.

Hyperscale construction runs through large national contractors, but the surrounding demand lands regionally. Projects like this pull on the whole industrial ecosystem around them: utilities scaling up power and water service, municipalities reviewing discharge permits and system capacity, and the manufacturers and facilities that grow alongside a build this size. That work needs control systems integration, process monitoring and instrumentation, commissioning, and long-term service.

For organizations already working in controls and automation in Ohio, a few things are worth watching:

1. Whether the lease and financing are finalized. The tenant list is apparently still in motion.
2. How the first phase tracks toward 2028. An 800-megawatt opening phase would by itself be one of the largest facilities in the region.
3. What it pulls on locally. Water capacity, discharge permitting, and grid coordination in southern Ohio will say a lot about how ready the region is.
4. Where your own capabilities overlap. The line between industrial systems and data center operations keeps blurring.

Ohio's position in this shift is still developing, but the direction is clear: the region is becoming a central node in AI infrastructure, and the disciplines that have run its plants for decades will matter more, not less. We have watched growth change the demands on water systems and industrial facilities across Ohio and Western Pennsylvania since 1968. When gigawatt-scale infrastructure arrives, the municipalities and facilities in its orbit feel it first, usually as new questions about monitoring, telemetry, and control systems. If those questions are surfacing at your plant, that is the work we do every day.

Seeing new demands on your water or industrial systems?

Talk to Steve, our Director of Sales Engineering. He helps municipalities and industrial facilities scope the monitoring, telemetry, and controls work that growth like this puts in motion.

Call+1(440) 708-1770

Emailsteveb@controlassociatesinc.com