Tech Giants Embrace Nuclear Power for AI Energy Demands

The escalating demands of artificial intelligence (AI) are driving major technology companies to reconsider and adopt nuclear power as a crucial energy source for their expanding data centers. Historically, concerns regarding cost and aging infrastructure had relegated nuclear energy to the background, but the immense power requirements of AI are now fostering a significant resurgence for this energy source.

According to the International Energy Agency (IEA), global data center electricity demand is projected to more than double by 2030, reaching approximately 945 terawatt-hours (TWh) compared to 415 TWh in 2024. Goldman Sachs analysts predict that the growth in AI demand will increase data center electricity needs by 160% by 2030. For example, OpenAI has stated the need for data centers requiring at least five gigawatts of energy. Hyperscale data centers, which are essential for AI processing, are now being designed to manage energy loads of up to 1 gigawatt (GW), a substantial increase from levels seen just a decade ago. This exponential growth necessitates a stable and high-capacity energy supply that intermittent renewables often struggle to consistently provide. Nuclear energy, characterized by its low-carbon emissions and high capacity factor, is emerging as a prominent solution. Currently, US nuclear power has a capacity of about 95 GW, accounting for approximately 19-20% of the country's total annual electricity generation, with a capacity factor of over 92.5%. The US Department of Energy projects that US electricity demand will grow by 15-20% by 2035, with data centers potentially consuming up to 9% of the total. The advancement and development of Small Modular Reactors (SMRs) and Micro Modular Reactors (MMRs) are further accelerating this trend, offering potential benefits such as faster licensing, reduced costs, and enhanced safety features.

A Deloitte report from April 2025 projects that new nuclear capacity could potentially meet approximately 10% of the projected rise in data center electricity demand by 2035, provided there is robust deployment over the coming decade. This projection highlights the growing recognition of nuclear energy's role in powering the future digital economy.

Major technology firms are actively integrating nuclear power into their energy strategies. In October 2024, Amazon announced a partnership with Dominion Energy and X-energy to develop over 5 GW of nuclear energy capacity by 2039, including an equity investment in X-energy and exploration of SMR development in Virginia. Simultaneously, Google revealed its collaboration with Kairos Power in October 2024, with plans to deploy up to seven SMRs, aiming for the first unit to be operational by 2030. This partnership is notable as the first corporate agreement for multiple deployments of an advanced reactor design in the U.S., intended to provide 500 MW of 24/7 carbon-free power. Additionally, Google has entered into an agreement with Elementl Power for the development of three advanced nuclear power plants, each with a capacity of 600 MW, the first of which is also expected to be commissioned by 2030. Google also plans to launch the Hermes 2 nuclear reactor in Tennessee in 2030, which will supply 50 MW of electricity for the company's data centers. Microsoft, in September 2024, entered into a 20-year agreement with Constellation Energy to resume operations at the Three Mile Island plant by 2028. will supply approximately 835 MW of power to Microsoft's data centers. Other technological giants, such as Apple, Meta, and Oracle, are also considering nuclear energy. Meta, for example, is exploring the possibility of adding one to four gigawatts of nuclear power by the early 2030s.

These strategic initiatives signify a notable industry-wide shift towards carbon-free energy solutions to meet the demands of the digital economy. According to BloombergNEF estimates, by 2035, AI-serving data centers could consume up to 4.4% of all global electricity, comparable to the energy consumption of countries like Germany or France. As the need for AI capabilities continues to surge, nuclear energy, particularly through SMRs and MMRs, is positioned to become a foundational component of future energy infrastructure.