The Department of Energy has identified 16 potential sites owned or managed by the DOE for rapid data center construction alongside in-place energy infrastructure. However, it is important that the accompanying energy infrastructure come from the most affordable, reliable sources rather than dictating winners and losers as a prerequisite for building data centers.
Secretary of Energy Chris Wright said in a prepared statement, “The global race for AI dominance is the next Manhattan project.” The Department of Energy intends for AI centers on these federal properties to commence operation by the end of 2027.
The 16 sites where AI infrastructure could be built are predominantly national laboratories, which offers “industry a chance to partner with DOE’s world-class research facilities.” The deals could include a combination of lease payments and an allocation of data center computing power to the colocated national laboratories.
The initiative builds off a Biden administration executive order that would lease public lands from the Department of Energy and Department of Defense for data centers—so long as the data centers are powered by “clean” energy. The executive order did not define if clean energy would be restricted only to wind turbines and solar panels, or if it would include carbon-free, baseload sources like nuclear energy. It will be paramount for the new Department of Energy to emphasize reliable energy—like natural gas, nuclear, and coal—and clearly define the energy generators that may be co-located with AI on these sites.
Companies are already recognizing the importance of securing reliable, affordable, and around-the-clock energy for data centers and procuring private purchase agreements with nuclear, natural gas, and other sources of baseload energy. Microsoft reached a deal to reopen Pennsylvania’s Three Mile Island nuclear plant by 2028.
Growing investment in data centers is expected to increase energy use, though there is considerable uncertainty given energy efficiency advancements. There has been “public pushback on data centers out of concerns customers will have to foot the bill for power plants.”
One legislative proposal in Minnesota is misguidedly attempting to require privately owned data centers to generate or procure at least 65% of their electricity from carbon-free sources and impose a fee on each megawatt of the data centers’ peak demand. The intent seems to be to insulate the electric grid from the consequences of growing demand and to dictate to data centers what generation technology to use. A recent Competitive Enterprise Institute study puts it best:
As the energy demands of AI grow, it is natural to encounter skepticism about prioritizing those needs, especially when weighed against potential environmental impacts. However, the role of the grid is not to dictate which industries deserve energy, but to efficiently supply power based on market demand, ensuring that innovation and economic growth are not constrained by artificial limitations. In short, the grid should meet consumer needs, not predetermine them.
The Department of Energy has requested information, seeking input from data center developers, energy developers, and the broader public. If the U.S. wishes to lead in artificial intelligence—and there are strong national security arguments that the U.S. should take this role, rather than China—it will need to avoid hampering data center development on private property, promote public-private partnerships on public lands, and ensure that all consumers have the affordable, reliable electricity they need.
This piece was originally published at Independent Women’s Forum Center for Energy and Conservation.