Letting Go in the AI-Nuclear Race

The Growing Demand for Power and the Role of Nuclear Energy

The rapid expansion of artificial intelligence (AI) has created an insatiable demand for power, particularly from data centers. This surge in energy consumption is putting pressure on existing infrastructure and prompting a search for sustainable and reliable energy sources. Among the potential solutions, nuclear energy stands out as a promising option due to its ability to provide consistent power over long periods. However, the United States is facing significant challenges in aligning its nuclear energy policies and funding with the fast-paced growth of AI.

The AI Explosion and Its Impact on Energy Needs

As of early 2024, there were over 10,000 data centers worldwide, with more than half located in the United States. These data centers are consuming vast amounts of electricity, with projections indicating that US data center power requirements could rise to 176 gigawatts (GW) by 2035, up from 33 GW in 2024. Companies like Amazon, OpenAI, and Meta are investing heavily in new data centers, which require massive amounts of energy to operate. For example, Amazon is constructing a large AI data center in Indiana that will become the state’s largest power consumer, relying primarily on natural gas for its energy needs.

Nuclear Energy as a Potential Solution

While current data centers rely heavily on fossil fuels, nuclear energy could offer a cleaner and more sustainable alternative. Existing nuclear reactors can be relicensed, uprated, or even restarted to meet the growing demand for power. Projects such as the reopening of Three Mile Island Unit 1 and the Palisades nuclear plant in Michigan demonstrate the potential for leveraging existing infrastructure. Additionally, companies like NextEra Energy are exploring the possibility of restarting mothballed reactors.

However, the development of new nuclear reactors remains slow. Westinghouse has announced plans to build ten new reactors in the United States, but these projects are not expected to begin until 2030, which may be too late to support the current AI boom. Smaller next-generation reactors are also being developed, but many are still in the design or licensing phase and may not be deployed for at least a decade.

Government Initiatives and Challenges

The Trump administration introduced a series of nuclear executive orders aimed at accelerating the deployment of nuclear energy. These policies include relicensing existing reactors, uprating their power output, and supporting the development of new reactor technologies. However, the effectiveness of these initiatives depends on a comprehensive strategy that includes financial incentives, regulatory reforms, and workforce development.

Despite these efforts, the U.S. government has not yet established a clear and robust framework for nuclear energy expansion. While the Department of Energy has identified four federal sites for potential AI data center infrastructure, the lack of a cohesive plan and sufficient funding remains a major obstacle. In contrast, countries like Russia and China have successfully leveraged government support to advance their nuclear programs, highlighting the importance of public-private partnerships.

The Need for Strategic Investment

To keep pace with the AI revolution, the United States must invest in a strategic package that includes expanded government financing, loan guarantees, and policy reforms. Without such measures, the country risks falling behind in the global race for nuclear energy leadership. The private sector alone cannot bear the burden of nuclear development, as demonstrated by the success of sovereign-backed projects in other nations.

In conclusion, while nuclear energy offers a viable solution to the growing power demands of AI, the U.S. must address its policy and funding gaps to ensure that it can compete effectively in this rapidly evolving landscape. The time to act is now, as continued delays could jeopardize the future of nuclear energy and the broader AI industry.