The rapid growth of AI applications has driven tech giants to explore nuclear energy as a sustainable power source for their expanding data centers. In addition to Oracle, Microsoft, and Google, Amazon Web Services (AWS) is also making significant strides in this area.
In 2023, the energy consumption of global data centers reached approximately 1-1.5% of total worldwide electricity use, a figure that is rapidly rising. To put this into perspective, Microsoft, which operates over 200 data centers globally, has seen its energy consumption skyrocket, driven by cloud services like Azure AI and AI-powered products such as OpenAI’s GPT models.
AI applications consume significant amounts of energy. For instance, a single request to OpenAI’s ChatGPT uses about 2.9 watt-hours of electricity. With billions of such queries daily, the energy demand is immense.
It’s been estimated that creating a single high-resolution image using DALL-E could consume several kilowatt-hours (kWh) of energy. Generating large-scale videos using AI—an increasingly popular service in media production—can use even more power, further straining the data centers that host such processes. In addition, the training phase for a large AI model, like GPT, can use the equivalent of hundreds of megawatt-hours (MWh) of electricity, comparable to the yearly energy consumption of hundreds of homes.
Data centers are critical to supporting AI applications, which require vast amounts of computational power. The International Energy Agency projects that data center electricity usage could exceed 1,000 terawatt-hours by 2026. This surge is driven by the increasing complexity of AI models and their integration across various sectors.
AWS announced plans to invest $650 million in a data center located next to the Susquehanna nuclear power station in Pennsylvania. This facility is expected to produce 2.5 gigawatts of power, underscoring AWS’s commitment to carbon-free energy23.
Oracle is moving forward with plans for a gigawatt-capacity data center powered by small modular reactors (SMRs). The company has already secured building permits for three reactors, highlighting its aggressive approach to meeting AI-driven energy demands.
Microsoft has partnered with Constellation Energy to reopen the Three Mile Island nuclear plant. This 20-year agreement aims to provide carbon-free energy for Microsoft’s data centers in the Pennsylvania-Jersey-Maryland region.
CEO Satya Nadella has also announced a partnership with Helion, a company developing fusion energy, with the goal of incorporating nuclear fusion into its power grid by the mid-2030s. Though nuclear fusion is still in its experimental phase, Microsoft is confident that this technology could revolutionize energy production, providing near-limitless, carbon-free power
Google is exploring partnerships with Kairos Power to develop small modular reactors. Although still in the planning stages, these initiatives reflect Google’s commitment to sustainable energy solutions.
Google’s data centers are already among the most energy-efficient in the world, and the company has been a pioneer in purchasing renewable energy. However, the scale of their AI operations and machine learning applications, including products like DeepMind and generative AI tools, has prompted a reconsideration of their energy mix
n the race to become the dominant AI platform, Oracle, Microsoft, and Google recognize that sustainability and energy security will be crucial competitive advantages. As they each invest heavily in nuclear power, they are positioning themselves to handle the increased demand for AI services without contributing to climate change. This not only aligns with their corporate sustainability goals but also allows them to offer AI services at scale with a significantly reduced carbon footprint.
Beyond competition in AI services, these companies are setting a standard for how the tech industry can lead the way in adopting cleaner and more efficient energy solutions. With AI projected to drive the future of multiple industries—ranging from healthcare to entertainment to manufacturing—those companies that successfully balance innovation with sustainability will be best positioned to lead the next wave of technological transformation.
The tech industry’s pivot towards nuclear energy highlights the urgent need for reliable and sustainable power sources to fuel AI advancements. Nuclear energy offers a promising solution, providing a stable and sustainable power source that can keep pace with the rapid expansion of AI-driven technologies. As companies like AWS, Oracle, Microsoft, and Google lead this charge, their investments in nuclear power could reshape the landscape of both technology and energy sectors.