Nuclear Power Hinges on Siting as AI-Driven Energy Demand Collides with Looming Gas Crisis

The rapid growth of artificial intelligence is putting a strain on the US electricity grid, and a looming natural gas crisis may soon exacerbate the issue.

Nuclear power could provide a reliable solution to this problem, but its adoption is being hindered by local opposition to new nuclear plant construction and data center development. According to Matthew Smith of Chronometer Partners, who spent 18 months building a detailed model of the US natural gas system, the country's traditional inventory cushion may be largely depleted by around 2030, leaving consumers vulnerable to price spikes.

Nuclear power offers a dependable source of carbon-free baseload electricity, with existing reactors providing well below average production costs compared to new gas-fired plants. Companies like Amazon and Microsoft are already sourcing power directly from nuclear facilities through long-term agreements, reducing their dependence on the broader electric grid and limiting the impact on residential ratepayers.

However, these companies face significant challenges in siting new data centers near existing nuclear plants, as residents often oppose AI and data center development, even if they support clean energy and economic growth in principle. This has led to a growing divide between those who favor expanding nuclear generation at existing reactor sites and those who prioritize building entirely new nuclear or data-center developments.