Welcome back to Voices of Tomorrow, the podcast where we explore the latest innovations shaping the future of AI, science, and beyond. In recent years, artificial intelligence has made leaps and bounds, revolutionizing industries, transforming societies, and even leading to two Nobel Prizes being awarded for AI-powered discoveries. But there’s another silent revolution happening behind the scenes—one that’s just as critical to the future of AI as the algorithms themselves. Today, we’re talking about the energy powering this AI revolution.

As AI models grow larger and more complex, the energy demands required to support them have skyrocketed. The same technology that powers language models, generative AI, and many other machine learning and computer vision algorithms is incredibly resource-intensive. And the traditional sources of renewable energy, like wind and solar, simply don’t provide the around-the-clock stability that AI operations require. The answer to this growing challenge? Nuclear energy. Yes, you heard that right—nuclear power, long seen as a legacy energy source, is now at the heart of AI’s future.

In this episode, we explore how tech giants like Google, Microsoft, and Amazon are harnessing nuclear energy to fuel their AI ambitions. Not only are these companies betting on nuclear power to provide the stability AI requires, but they’re also pioneering a new path forward for clean, reliable energy in the tech industry.

Let’s dive in.

Our story begins with Google, a company that has long been at the forefront of technological innovation. Google recently announced that it will back the construction of seven small nuclear reactors in the U.S., designed specifically to provide clean, round-the-clock power for its AI data centers. This move marks the first time a major tech company has fully committed to nuclear energy, and it’s a significant milestone in both the tech and energy industries.

The reactors, developed by Kairos Power, will supply up to 500 megawatts of power, enough to fuel an entire AI data-center campus. What makes these reactors special is their design. Unlike traditional nuclear power plants, which are massive, costly, and take decades to build, these are small modular reactors—far more flexible, cost-effective, and capable of being constructed in a fraction of the time.

So, why nuclear, and why now? Simply put, AI demands a lot of power. As Google and other AI companies like Microsoft and Amazon continue to push the boundaries of AI development, its data centers have become increasingly energy-hungry. Traditional renewable sources like wind and solar, though important, aren’t always reliable. Nuclear energy, on the other hand, offers a stable, carbon-free power source that’s perfect for fueling continuous AI operations.

This partnership with Kairos Power is more than just a power deal—it’s a signal of where the future of AI and energy are headed. Google’s as well as Microsoft’s and Amazon’s bold move into nuclear is setting the stage for other tech companies to follow, demonstrating that the key to sustaining the AI revolution may lie in embracing clean, innovative energy sources like nuclear power.

But who is Kairos Power? Kairos Power is a trailblazer in the nuclear energy sector that’s pushing the boundaries with its innovative molten salt–cooled reactors. Kairos’ reactors are designed not just to power today’s AI needs but to redefine the way nuclear energy is approached altogether.

At the heart of Kairos’ reactor design is the use of a molten salt coolant called Flibe, which is a mixture of fluorine, lithium, and beryllium. Unlike traditional water-cooled reactors, which require high-pressure systems, Kairos' molten salt design operates at atmospheric pressure, drastically reducing risks associated with high-pressure systems. This inherently safer design allows for a simpler, more efficient operation, which can provide clean, round-the-clock energy at costs competitive with natural gas.
Additionally, Kairos employs TRISO fuel, a cutting-edge fuel composed of tiny uranium kernels encased in multiple protective layers. These layers make the fuel resistant to high temperatures and other safety concerns, even under extreme conditions. The end result is a nuclear power source that is safer, more reliable, and ideally suited for supporting the continuous energy demands of AI data centers.

Kairos’ reactors offer more than just technical innovation—they signal a shift in how nuclear power is perceived and applied. With plans to build a demonstration reactor, Hermes, by 2026, Kairos is on track to revolutionize both the nuclear industry and the tech sector by making nuclear energy a key player in the AI revolution. Do not confuse the Hermes reactor with Hermès International, the French luxury brand established in 1837, although both are equally unique and high-quality.

On a more serious note, it is important to note that Google isn’t the only tech giant turning to nuclear energy to power its AI infrastructure. As already mentioned, Microsoft and Amazon are also making significant investments in nuclear power, recognizing that the future of AI depends on securing a stable, scalable energy source.

Microsoft, in a historic move, has partnered with Constellation Energy to revive a nuclear reactor at Pennsylvania’s Three Mile Island. This deal involves restarting the undamaged reactor to supply clean energy to Microsoft’s cloud and AI data centers. The newly named "Crane Clean Energy Center" will provide the round-the-clock power Microsoft needs to fuel its growing AI workloads over the next two decades.

Amazon, too, is making waves by securing nuclear power to drive its AI-powered cloud services. Earlier this year, Amazon Web Services (AWS) acquired a data center near Pennsylvania’s Susquehanna nuclear plant, marking its first-ever foray into nuclear energy. This move reflects Amazon’s recognition that AI’s future energy demands cannot be met by traditional renewables alone. Nuclear, with its capacity for continuous power generation, is perfectly positioned to keep Amazon’s cloud services running smoothly, ensuring that AI operations are never interrupted by power shortages or grid fluctuations.

These strategic investments by Microsoft and Amazon represent more than just energy deals—they are a clear indication that nuclear energy will play a pivotal role in the future of AI. As AI continues to evolve, its energy requirements will only grow, and the need for clean, reliable, and scalable power sources will become more urgent. Nuclear energy provides a solution to this challenge, offering the tech industry a way to power the future without compromising on sustainability or reliability.

As we’ve seen today, the future of AI is deeply intertwined with the future of energy. The rapid advancements in AI technology—from language models to image analysis—have opened the door to incredible possibilities, but they have also presented new challenges, particularly in the form of energy consumption. And now, tech companies are looking to nuclear energy not just as a backup plan, but as a cornerstone of their AI infrastructure.

Google, Microsoft, and Amazon are leading the charge, demonstrating that nuclear power—long sidelined due to public perception and high costs—can be the key to sustaining the AI revolution. By adopting nuclear energy, these companies are not only meeting their energy needs but also paving the way for a cleaner, more sustainable future.

Nuclear energy is no longer a relic of the past. It’s a vital part of the solution for the tech-driven world of tomorrow. As AI continues to grow, so too will the demand for energy solutions that are both scalable and environmentally friendly. And nuclear energy, with its ability to provide constant, carbon-free power, is perfectly suited to meet this demand.

In today’s episode, we’ve explored how nuclear energy is becoming the backbone of the AI revolution, powering the massive data centers driving advancements in artificial intelligence. This need for scalable energy sources ties back to our discussion in Episode 5: Models of Tomorrow: Scaling Laws in Machine Learning, where we examined how the compute demands of AI models scale exponentially with size. As AI systems grow larger and more complex, their energy requirements follow similar scaling laws, making reliable, high-output energy sources like nuclear power essential to sustaining this progress. 

Today’s episode also connects seamlessly with Episode 6: The Dark Matter of Biology: How AI is Mapping the Hidden RNA Virosphere, where we discussed how AI is unlocking the secrets of the RNA virosphere. Just as nuclear power is fueling the scaling of AI, the AI systems themselves are fueling breakthroughs in fields like biology. The computational power needed to analyze massive datasets, like those used to uncover new RNA viruses, requires immense, stable energy sources. Nuclear energy ensures that AI can continue to push the boundaries of discovery, whether in biology, physics, or any other frontier.

Thank you for tuning into Voices of Tomorrow. We hope today’s discussion has sparked new ideas and insights into the role nuclear energy will play in powering the future of AI. As always, your thoughts and feedback are invaluable, so be sure to share them with us. Together, we’re exploring the innovations shaping tomorrow, and there’s no limit to what we can achieve when we combine the power of AI with the resources of tomorrow. Stay inspired, and keep looking toward the future!