Key Takeaways
- Data centers take years to build but contracts come early. Curious hosting providers may want to follow where AI infrastructure is sprouting up next.
- Northern Virginia and Texas dominate, but Ohio and a couple others now support more than 1 GW of AI data center capacity each.
- Major hubs offer the infrastructure, while growing markets offer room to grow. So, which road should providers take?
Northern Virginia and Texas have long been major data center markets, but AI infrastructure is officially doing what we all suspected it would eventually do: spread into more states.
A new study from data infrastructure provider TRG Datacenters mapped the U.S. states best positioned for AI infrastructure growth based on existing capacity, electricity use, emissions, and water stress.
Ohio, Tennessee, Indiana, and Mississippi are already seeing impressive AI infrastructure development, which means there are a few new names worth adding to the expansion watchlist.
AI Infrastructure Is Expanding Into New Markets
The U.S. data center map is well on its way to going beyond a few hubs spread throughout the coastal states from east to west. (And yes, I’m counting Ohio as landlocked.)
To be clear, the study doesn’t measure hosting demand, but it does show where AI infrastructure is developing. Mississippi, South Carolina, and Oregon, for example, all appear in the study alongside the more well-known markets, like Virginia.
In fact, Virginia topped TRG’s index. But some of the biggest capacity numbers actually came from states just outside Data Center Alley: Ohio has 1,410 megawatts (MW) of AI data center capacity, followed by Tennessee at 1,373 MW.
Where Should Hosting Providers Look Next?
Don’t look at TRG’s list and immediately start picking states off like a shopping list. It all comes down to what you’re looking for. Some states already have huge infrastructure footprints, while others are building enough capacity to eventually become serious prospects.
AI Data Center Capacity Across the U.S.
Source: TRG Datacenters
Here’s how I’d break it down:
Established Giants: Virginia and Texas
Let’s start with the obvious ones. Virginia and Texas are already major data center markets: Virginia has seven large AI data center sites and more than 1 GW of capacity. Texas has 13 large sites, which is the most in the entire study.
The infrastructure is already there, which makes these markets an obvious place for hosting providers looking to expand. Of course, the downside is that these places aren’t exactly a hidden gem.
These are crowded, highly competitive markets, so actually finding available space, power, and capacity can be a challenge. For providers already operating there, expanding may be a natural move. For everyone else, it may be worth looking at what’s developing outside the usual suspects.
Up-and-Coming Large-Capacity Markets: Ohio, Tennessee, and Indiana
Ohio, Tennessee, and Indiana obviously don’t have nearly as many large AI data center sites as Texas. And yet, they’re already supporting some serious capacity: Ohio has 1,410 MW, Tennessee has 1,373 MW, and Indiana has 1,088 MW, according to TRG’s findings.
Ohio is particularly interesting because its AI capacity is actually higher than Virginia’s despite having fewer large sites. Tennessee tells a similar story, with just three large sites accounting for more than 1.3 GW of capacity.
We’re looking at states that are already attracting large amounts of computing capacity, even if they don’t have the same number of facilities as the traditional giants. After all, infrastructure tends to attract more infrastructure.
Developing Alternatives: Georgia, Mississippi, South Carolina, Oregon, and Others
Georgia has two large AI sites and 664 MW of capacity. Mississippi has two sites and 512 MW, while South Carolina and Oregon each have one large site.
Yes, they’re smaller than the established markets, but they’re still growing outside the usual data center hubs, which is good because the biggest data center markets are already getting crowded.
Northern Virginia, for example, added more than 1 GW of data center capacity over the past year, yet CBRE found that its vacancy rate had fallen to just 0.2%, leaving only 10.8 MW available. It also found that across eight major North American markets, 80.4% of data center capacity under construction had already been preleased.
Which Way Should Providers Go?
With AI sending demand for data centers skyrocketing and builders struggling to keep up, you don’t really have to look at the biggest markets to get a sense. If you’re planning to expand or add more servers, you can follow the lead, so to speak, and look at where the investors are, well, investing.
Industrial electricity prices vary quite a bit across the country as well. Through June 2026, the average industrial electricity price was 6.65 cents per kilowatt-hour in Texas and 10.08 cents in Virginia, according to the U.S. Energy Information Administration. Tennessee was even lower at 6.77 cents, while Ohio came in at 10.21 cents.
So which road is better? Maybe neither. Maybe both.
Established markets like Virginia and Texas already have the infrastructure to support large-scale data centers, but available capacity is getting harder to come by than it used to. On the other hand, newer markets can offer more room to grow and maybe even lower costs if you get in early. Of course, you still need the basics: power, land, fiber, permits, and a grid connection.
Either way, data centers take years to build, so if you know you’re going to need more server space or compute capacity, it probably makes sense to start looking before you actually need it.




