Small Nuclear Reactors: A Gamble to Power AI's Surging Energy Demands

The AI Power Crunch and the Nuclear Solution

As artificial intelligence (AI) continues to revolutionize various sectors, the technology's voracious appetite for electricity is becoming increasingly apparent. Data centers, the backbone of AI infrastructure, are projected to account for 9% of the entire U.S. electricity demand by 2035, marking a significant 5% increase from current levels 1

1

. This surge in power demand has led tech giants and investors to explore innovative solutions, with small modular reactors (SMRs) emerging as a potential answer.

The Rise of Small Modular Reactors

SMRs are compact nuclear reactors designed to generate between 300-500 megawatts of power, significantly smaller than traditional reactors that typically start at 1,000 megawatts 1

1

. Their modular design allows for easier construction, potentially reducing costs and build times compared to conventional nuclear plants. Tech companies are showing keen interest in this technology, with Google partnering with Kairos Power to develop 500 megawatts of small nuclear reactors by the mid-2030s, and Equinix collaborating with Oklo, backed by OpenAI's Sam Altman 1

1

.

Global Interest and Investment

The enthusiasm for SMRs extends beyond the United States. The UK government is investing £2.5 billion ($3.4 billion) in SMR development over the next decade, while France's state-owned Electricité de France SA is backing the Nuward SMR project 1

1

. Russia and China have already launched operational SMRs, with capacities of 35 and 100 megawatts, respectively 1

1

.

Economic Viability and Challenges

Despite the growing interest, experts warn that the economic viability of SMRs remains uncertain. The technology has attracted over $9 billion in investments since 2019, with the U.S. government committing more than $6 billion and private investors contributing over $3 billion 3

3

. However, past SMR projects have faced significant delays and cost overruns, raising concerns about their competitiveness with other power sources.

Cost Comparisons and Skepticism

Estimates of SMRs' levelized cost of energy vary widely. Wood Mackenzie projects that by 2030, SMRs could generate power at $182 per megawatt-hour, compared to $133 for conventional nuclear and $126 for natural gas 3

3

. Some SMR developers claim much lower costs, with Oklo and NuScale projecting $90 and $64 per megawatt-hour, respectively 3

3

. However, nuclear engineers argue that larger reactors may be more cost-effective due to economies of scale.

Supply Chain and Fuel Concerns

The SMR industry faces additional challenges related to fuel supply. Most SMRs are designed to use high-assay low-enriched uranium, which is primarily produced in Russia. With limited production capacity in the U.S., companies may need to rely on government stockpiles, potentially adding up to $20 per megawatt-hour to power costs 3

3

.

The Path Forward

Despite the challenges, proponents of SMRs remain optimistic about the technology's potential to provide clean, reliable power for AI data centers and contribute to climate change mitigation efforts 2

2

. As the AI boom continues to drive electricity demand, the success of SMRs could play a crucial role in shaping the future of both the tech and energy sectors.