As the demand for artificial intelligence (AI) technologies skyrockets, tech giants are searching for innovative ways to meet the enormous energy requirements that come with it. Google, a global leader in AI and cloud computing, has recently taken a bold step by signing a deal with nuclear power company Kairos Power to provide clean, reliable energy to its data centers. This agreement marks a significant shift in how technology firms address energy challenges, with advanced nuclear reactors playing a central role in powering the future of AI.
The deal with Kairos Power involves using small modular reactors (SMRs) to generate electricity for Google’s vast network of data centers, which house the servers that drive its AI systems. These reactors, which are smaller and more efficient than traditional nuclear power plants, represent a cutting-edge approach to energy generation. Google’s commitment to bringing the first reactor online within this decade, with additional reactors by 2035, signals its intent to lead the way in decarbonizing the tech industry’s energy supply.
The Energy Challenge of AI
Artificial intelligence has revolutionized industries worldwide, from healthcare and finance to retail and entertainment. AI’s ability to process large datasets, generate insights, and improve decision-making has positioned it as a key driver of economic growth. However, this technological revolution comes with a significant energy cost. AI data centers, which operate 24/7, require vast amounts of electricity to run advanced computing hardware and cooling systems. Without reliable energy sources, the growth of AI could be hampered by power shortages or environmental concerns.
In a statement, Michael Terrell, Senior Director for Energy and Climate at Google, acknowledged the challenge, saying, “The grid needs new electricity sources to support AI technologies. This agreement helps accelerate a new technology to meet energy needs cleanly and reliably, and unlock the full potential of AI for everyone.”
Global energy consumption by data centers is expected to more than double by 2030, according to a report by Goldman Sachs. AI workloads, especially those involving machine learning and neural networks, demand more energy than traditional computing tasks. These data centers are filled with energy-intensive hardware, including specialized chips like graphics processing units (GPUs) and tensor processing units (TPUs), which require constant power and generate significant heat.
John Moore, Industry Editor for the TechTarget website, highlighted the unique energy demands of AI. He told the BBC, “These data centers are equipped with specialized hardware… that require lots of power, that generate lots of heat.” Traditional energy sources, including fossil fuels and renewables, may struggle to keep up with this rapid growth, leading companies like Google to explore alternatives like nuclear power.
Nuclear Power: A Viable Solution
Nuclear energy offers a promising solution to the growing energy demands of AI technologies. Unlike fossil fuels, nuclear power is virtually carbon-free, making it a sustainable option for companies looking to reduce their environmental impact. Additionally, nuclear plants provide a consistent energy supply, operating around the clock without the intermittency issues associated with wind or solar power.
Jeff Olson, an executive at Kairos Power, emphasized the importance of this partnership in advancing nuclear technology. “The deal with Google is important to accelerate the commercialization of advanced nuclear energy by demonstrating the technical and market viability of a solution critical to decarbonizing power grids,” he said.
One of the key innovations in Kairos Power’s approach is the use of molten fluoride salt as a coolant, rather than water. Traditional nuclear reactors rely on water to cool their reactors, which can create challenges such as the need for large bodies of water and the risk of overheating. By contrast, molten salt reactors are more efficient, safer, and better suited to the needs of smaller, modular systems like those envisioned by Kairos Power.
In July 2023, the company began construction on a demonstration reactor in Tennessee, marking the first step toward the commercial rollout of its technology. While the timeline for bringing these reactors online is ambitious, the potential benefits are enormous. By adopting nuclear power, Google is positioning itself at the forefront of a new era in energy technology—one that could transform how the tech industry powers its operations.
The Broader Shift Toward Nuclear Energy
Google is not alone in recognizing the potential of nuclear power to address the energy demands of AI and other technologies. In recent months, several major tech companies have made similar moves. In March 2024, Amazon announced plans to acquire a nuclear-powered data center in Pennsylvania, marking its first foray into nuclear energy. Microsoft also made headlines by securing a deal to restart operations at the Three Mile Island energy plant, which was the site of the United States’ worst nuclear accident in 1979.
These moves are part of a broader trend toward embracing nuclear power as a key component of decarbonization efforts. At the United Nations Climate Change Conference in 2023, the U.S. joined a coalition of countries aiming to triple their nuclear energy capacity by 2050. This ambitious goal reflects the growing recognition that nuclear power, despite its challenges, offers a viable path toward reducing carbon emissions and combating climate change.
However, the use of nuclear energy is not without controversy. Critics point to the risks associated with nuclear power, including the potential for accidents and the long-term challenge of managing radioactive waste. High-profile disasters like Chernobyl and Fukushima have left a lasting legacy of fear around nuclear energy, and public perception remains a significant hurdle to its widespread adoption.
Yet, supporters argue that advances in technology, such as those developed by Kairos Power, have made modern nuclear reactors safer and more efficient than ever before. Small modular reactors, in particular, are designed with enhanced safety features that reduce the likelihood of catastrophic accidents. Moreover, the use of molten salt as a coolant offers significant advantages over traditional water-cooled reactors, making it a more sustainable and manageable option for long-term energy generation.
Regulatory Hurdles
Before Google’s nuclear-powered data centers become a reality, the project must clear several regulatory hurdles. The U.S. Nuclear Regulatory Commission (NRC) and local agencies will need to approve the plans, ensuring that safety standards are met and that the reactors comply with environmental regulations. Given the history of public concern around nuclear power, this approval process could take time.
Last year, Kairos Power made significant progress when it received the first permit in 50 years to build a new type of nuclear reactor in the United States. This milestone suggests that regulatory bodies are increasingly open to the idea of advanced nuclear technologies, but challenges remain.
Conclusion
Google’s decision to turn to nuclear power for its AI data centers marks a significant development in the ongoing quest to meet the energy demands of the digital age. By partnering with Kairos Power and embracing small modular reactors, the company is not only addressing its immediate energy needs but also setting a precedent for the tech industry. As the world moves toward a future powered by AI, nuclear energy may prove to be the key to ensuring that this technological revolution is sustainable, reliable, and environmentally responsible.
