Big tech giants including companies including Amazon (NASDAQ: AMZN), Google-parent company Alphabet (NASDAQ: GOOGL), and Microsoft Corporation (NASDAQ: MSFT) are exploring nuclear energy as a potential power source for AI development.
Artificial intelligence, and in particular, the training of large language models and running data centers, requires an immense amount of electricity.
Traditional energy sources and renewable sources like solar and wind alike struggle with consistency and capacity issues. Here, nuclear energy presents a compelling case with its ability to provide a steady, high-density power supply with a significantly lower carbon footprint compared to fossil fuels.
Google CEO Sundar Pichai said the company aims to use nuclear energy to power its growing AI data center network.
The CEO said Google is exploring small modular nuclear reactors to help meet its ambitious climate goals. Google aims to achieve net-zero emissions across its operations by 2030, a target shared by other Big Tech companies. Microsoft has pledged to be carbon negative by 2030, while Amazon has committed to reaching carbon neutrality by 2040, with a promise to use 100 per cent renewable energy by 2025.
Despite these goals, some critics argue that Big Tech’s climate targets are unrealistic. In 2023, Google’s total greenhouse gas emissions increased by 48 per cent compared to 2019, driven by investments in energy-intensive GenAI.
Except that nuclear energy as a low-carbon, high-output solution to meet the large-scale power demands of AI infrastructure. Unlike fossil fuels, which contribute significantly to global greenhouse gas emissions, nuclear power generates electricity without releasing carbon dioxide during operation.
Read more: Palladyne AI flies high after completing first phase of U.S. Air Force contract
Read more: Joby Aviation soars after Toyota agrees to invest half a billion
Nuclear energy is efficient because of energy density
Nuclear power offers several key advantages for tech giants.
It provides reliable energy, as it isn’t dependent on weather conditions like renewable sources, ensuring a consistent power supply critical for the uninterrupted operation of data centers.
Nuclear energy is also highly efficient due to its energy density, with nuclear reactions generating vast amounts of energy from a small amount of fuel. This efficiency supports the goal of companies striving for net-zero emissions, as nuclear power produces minimal greenhouse gases during operation.
Additionally, with the right infrastructure, nuclear energy can be scaled to meet growing demands, which is increasingly important as AI’s energy consumption continues to rise rapidly.
Furthermore, recent discussions and deals highlight a trend of companies like Microsoft and Google actively pursuing partnerships in nuclear energy. Microsoft’s agreement with Constellation Energy to potentially reactivate a nuclear reactor at Three Mile Island marks a bold step into this area. Likewise, Google’s focus on atomic power for its data centers emphasizes the industry’s shift toward nuclear energy solutions.
Integrating nuclear power into Big Tech’s energy mix presents several challenges.
Strict regulations govern the nuclear industry, requiring companies to navigate a complex landscape of safety protocols, public perception, and the historical impact of nuclear accidents.
Additionally, building new reactors or restarting old ones involves significant time and expense, as seen in the Vogtle plant expansion in Georgia, where delays and cost overruns have been substantial.
The issue of nuclear waste disposal also remains contentious, requiring solutions that are both technically sound and acceptable to the public.
Read more: Westinghouse Electric Company inks deal to provide UK first privately-financed SMR fleet
Read more: Westinghouse Electric partners with Saskatchewan’s research arm to build nuclear microreactor
Nuclear plants “safest and most secure facilities”
A May 2024 report from Australia’s CSIRO revealed that nuclear power from both large-scale reactors and small modular reactors is significantly more expensive than electricity generated from renewables.
The report estimates that large-scale nuclear costs between USD$141 and USD$233 per megawatt hour, while small modular reactors range from USD$230 to USD$382 per megawatt hour. In contrast, electricity from solar and wind would cost only USD$73 to USD$128 per megawatt hour.
Nuclear advocate, the International Atomic Energy Agency asserts that nuclear plants are among “the safest and most secure facilities in the world.”
Nuclear plants undergo rigorous safety testing and must comply with international safety standards. However, high-profile accidents have still occurred, releasing significant amounts of radioactive material into the environment.
In 1986, the Chernobyl nuclear power plant in Ukraine experienced the worst nuclear accident in history, releasing massive radiation across Europe. More recently, in 2011, a tsunami flooded the Fukushima nuclear plant, causing the meltdown of three reactor cores. Although fatalities from radiation exposure were minimal, the accident caused widespread environmental damage.
The fear of such disasters makes nuclear energy a contentious issue, but it is important to note that Chernobyl and Fukushima were caused by flawed designs or natural disasters. Newer nuclear technologies aim to reduce these risks with improved safety features.
The intersection of AI and nuclear energy could herald a new phase in technology’s evolution. Leaders like Jensen Huang of Nvidia have publicly emphasized the necessity of nuclear power for advancing AI, suggesting a future where technological growth directly depends on advancements in energy production.
.