OPINION:
When xAI’s Colossus data center opened near Memphis, Tennessee, in July 2024, it began drawing roughly 150 megawatts continuously from the grid, about as much electricity as a midsize city. It supplemented that with substantial on-site generation.
Colossus is one of thousands of data centers now planned or in development across the United States. BloombergNEF estimates that U.S. data center demand could reach 106 gigawatts by 2035.
For decades, new power generation had to keep pace only with retirements; it didn’t need to anticipate load growth. In that environment, steady increases from grid-tied intermittent renewables, alongside natural gas, were enough to push down coal’s share of the power mix from nearly 50% in 2010 to just 16% today.
As demand accelerates, the gap between intermittent supply and firm need will reveal itself in higher prices and tighter reliability margins.
The U.S. has a few options. Mitigated natural gas holds promise, as does advanced geothermal. Both deserve continued development. These technologies require additional commercial proving, durable political support and significant capital formation.
Beyond 2045, their contribution may be substantial, but that’s too late to end-run the impending economic, environmental and social pain the U.S. economy will undergo if power supply conditions don’t improve over the next 20 years.
Nuclear is the only proven clean energy source that can deliver firm power at scale in the 2030s and 2040s. Uprates at existing plants and a handful of reactor restarts can add incremental capacity quickly, but the inventory is limited.
To meaningfully close the gap, the U.S. needs greenfield development: large reactors such as the AP-1000s in Georgia and advanced designs now moving toward commercial demonstration. The Department of Energy’s Office of Energy Dominance Financing is already working with utilities to underwrite perhaps a dozen reactors with targeted go-live dates from 2031 to 2035.
After a long hiatus in large-scale reactor construction, first-of-a-kind costs remain high and the financial sector is understandably wary of cost overruns. Without a mechanism to share that early risk, projects stall and the gap grows.
Sens. Jim Risch, Idaho Republican, and Ruben Gallego, Arizona Democrat, recently introduced the ARC Act, a targeted mechanism to share first-of-a-kind cost risk. ARC is not a sweeping industrial policy bill; it’s a pragmatic market development tool.
ARC is more impressive when considering the downstream supply chain effects. Establishing design and execution know-how, reinvigorating the nuclear construction workforce and activating teams of financial risk analysts and underwriters would carry real weight for the next dozen reactors, and costs would fall.
ARC would merit support simply for its pragmatic, fiscally modest qualities, but its value runs deeper than that. Many popular clean energy policies, such as renewable portfolio standards, effectively socialize costs through retail electricity rates. It’s hard to imagine a more regressive public finance model. In contrast, ARC is a targeted federal risk-sharing mechanism that spreads costs more equitably and ties public exposure directly to economic outcomes.
There isn’t much time, and inaction will come at a cost. Power prices will continue to increase because the U.S. spent a bit too long cultivating an underdiversified generation mix.
The U.S. can’t rely only on legacy-defining industrial policy and demand-pull megabills such as those that brought intermittent renewables from calculators and satellites to cost-competitive generators in less than two decades. It must beat a consistent drum of cogently staged, unsentimental market development policies to open the throttle for nuclear and other clean, firm generation sources on the soonest available timeline.
The ARC Act is exactly that kind of policy. ARC places a modest federal Treasury bet on an investment that could pay for itself relatively quickly. Lower power prices would support broader economic activity and, in turn, higher federal tax receipts.
It’s a standout example of the strategically targeted, bipartisan energy policy we need more of to keep costs from soaring.
• Erik Funkhouser is the executive director at Good Energy Collective. Julia Sweatman is a senior policy associate at Good Energy Collective.
Please read our comment policy before commenting.