Commonwealth https://www.cai-engr.com 245 West Michigan Avenue,
Jackson, MI 49201
517.788.3000

Apr 21, 2025

The U.S. Department of Energy has publicly identified more than a dozen federal sites that could potentially be opened to companies willing to build new data centers.



In a somewhat surprising move, the U.S. Department of Energy (DOE) has publicly identified more than a dozen federal sites that could potentially be opened to companies willing to build new data centers. The announcement comes on the heels of the current administration's commitment to supporting certain policies enacted by the previous administration.

If the DOE follows through, it is possible we could see new data centers springing up at revered sites like Los Alamos, the storied federal laboratory that played a monumental role in developing the atom bomb. The Oak Ridge and Sandia federal laboratories have also been mentioned as possible data center sites.


The Sites Are Ready to Go


All the sites on the government's list are apparently ready to go. The national laboratories were chosen specifically because they are "uniquely positioned for rapid data center construction, including in-place energy infrastructure with the ability to fast-track permitting for new energy generation such as nuclear [power]."


As you may already know, data centers are power-hungry facilities. They consume a tremendous amount of power for both data storage and the mathematical computations responsible for keeping computers and servers running. But data centers also consume power to keep servers cool. Any location suitable for a new data center needs to have access to power in some way, shape, or form.



Washington's Motivation to Get Involved


As for Washington, why would regulators want to get involved? It is all about artificial intelligence (AI). Several years ago, then-president Joe Biden made it clear that he wanted the U.S. to be a world leader in developing AI technology. But in order to do so, more data centers need to be built.


AI requires tremendous computational resources. It requires power-hungry CPUs along with enough servers to gather, store, and analyze unfathomable amounts of data. Running at all requires power – and a lot of it.


Under the Biden administration, the goal was to fast-track the combination of new data center construction and green power generation. The administration's vision included new data centers powered by nearby plants built specifically for that purpose and relying on sustainable technologies.


Surprisingly, the Trump administration shares this particular goal with its predecessor. But the current administration wants to take things one step further. By identifying federal sites that could host both data centers and power plants, the opportunity to move very quickly is at hand.



Getting Data Centers Online More Quickly


Federal sites are largely exempt from state regulations. Therefore, with a cooperative administration willing to cut red tape and fast-track the permitting process, data centers could be built and brought online more quickly. In addition, the current administration is open to a variety of power generation solutions, including nuclear.


Getting data centers online as quickly as possible is seen as a critical component in taking the lead in AI development. Other countries are building their own data centers and power plants. If we don't keep up, we will likely be surpassed in the AI world. Neither the current nor previous administration is okay with that.


Now that the sites are all but identified, we wait to see if anyone bites. Knowing what we know about the power generation industry and its relationship to data centers, few would be surprised if companies began lining up to get their shot at the sites.


When all is said and done, the sixteen identified sites might not be enough. Either way, bringing more than a dozen new data centers online could certainly make a difference in the race for dominance in AI development and production.

Left arrowBack to All Insights

Related Expertise

Related Markets