The New Arms Race: AI, Energy, and the Return of Nuclear Power

In a high-level meeting in Washington this past week, the CEOs of some of the world’s largest energy companies convened with senior U.S. government officials and representatives of the technology sector to address a quietly mounting crisis: how to meet the vast and accelerating energy demands driven by artificial intelligence (AI).

At the heart of their discussions lies a striking projection: global data center capacity is expected to triple within five years. Roughly one-third of this expansion will occur in the United States alone, requiring up to 150 gigawatts of additional power capacity—equivalent to the combined electricity needs of ten cities the size of New York. This is not a theoretical scenario. The exponential rise of AI model training, cloud computing, and real-time applications is already placing immense strain on existing grids.

And yet, this digital explosion is occurring atop infrastructure that was largely designed for the demands of the last century. Outdated grid systems, slow permitting procedures, and fragmented energy policy frameworks are proving insufficient. The result is a growing gap between the electricity required to power AI and what the system can actually deliver.

As Chris Lehane, OpenAI’s chief global affairs officer,  aptly underscored, the capacity to rapidly scale reliable, high-density energy infrastructure has become a defining factor in global technological leadership. At present, only two countries possess the financial, industrial, and regulatory apparatus to pursue this at the required pace: the United States and China. China’s recent commissioning of ten new nuclear power plants in a single year is a case in point—underscoring how energy policy has now become inseparable from geopolitical strategy.

The World Bank’s Return to Nuclear Energy

In parallel with these developments, a landmark policy shift is underway. After more than six decades of near-complete disengagement, the World Bank has announced that it will re-enter the nuclear energy space. This decision marks a decisive turn in international development finance.

The Bank estimates that by 2035, electricity demand in developing economies will double, pushing annual infrastructure investment needs from $280 billion today to $630 billion. Meeting this demand while maintaining progress on climate goals has compelled a reassessment of nuclear power’s role in the global energy mix.

For the first time in decades, nuclear energy is being viewed not merely as a technological option but as an indispensable part of a balanced, decarbonized, and resilient electricity portfolio. Industry bodies, such as the World Nuclear Association and the International Atomic Energy Agency (IAEA), welcomed the decision as a long-overdue recognition of nuclear’s unique contribution to clean energy goals.

The World Bank’s initial focus will be conservative and targeted: supporting life-extension projects for existing reactors and accelerating the deployment of Small Modular Reactors (SMRs)—innovative technologies offering lower capital requirements, faster construction timelines, and greater deployment flexibility. Although proliferation concerns persist, the Bank is coordinating closely with the IAEA to develop robust safeguards, compliance mechanisms, and internal expertise before scaling further.

It is notable that this policy pivot was made possible, in part, by a shift in Germany’s position on nuclear energy, as well as a growing bipartisan consensus in the United States that views nuclear as vital to both climate resilience and national competitiveness.

Nuclear Power as the Backbone of AI-Driven Infrastructure

There is a fundamental truth underpinning the AI revolution: AI is not powered by code alone—it is powered by electricity. The development and deployment of AI systems require immense and sustained computational resources. Training large-scale models often consumes the output of entire power plants over the course of weeks or months. And once deployed, these models must remain constantly active—processing queries, generating responses, and integrating across digital systems in real time.

This shift is creating unprecedented demand for baseload power—electricity that is available around the clock, irrespective of weather conditions or peak load cycles. Renewable energy sources, while essential to the broader transition, remain inherently intermittent and require large-scale storage solutions to deliver consistency. Fossil fuels, while dispatchable, are incompatible with net-zero commitments and global climate objectives.

Nuclear energy is uniquely suited to bridge this divide. It offers carbon-free, high-capacity, stable generation that can operate continuously and reliably. For data centers and AI infrastructure—where uptime, predictability, and scalability are non-negotiable—nuclear is not a fallback option; it is a critical enabler.

Beyond its technical characteristics, nuclear energy also delivers strategic resilience. It reduces exposure to volatile fuel markets, geopolitical shocks, and supply chain vulnerabilities—factors that are increasingly relevant as AI becomes integrated into national defense, health systems, and economic infrastructure.

In this sense, nuclear power is no longer a matter of energy policy alone. It is a cornerstone of technological sovereignty and competitive endurance. As China accelerates its twin investments in AI and nuclear capacity, the West faces a strategic imperative: to build an energy ecosystem capable of sustaining its digital ambitions.

Conclusion

Artificial intelligence is reshaping every domain it touches—from law and medicine to logistics and warfare. But its progress is tethered to an old-world constraint: the availability of clean, reliable, scalable power. The real competition, therefore, is not just for better algorithms—but for the infrastructure that makes them possible.

The revival of nuclear energy—both in policy and in investment—is not simply a return to a legacy technology. It is a forward-looking response to the energy realities of a digital future. For lawyers, policymakers, and investors alike, understanding this convergence of AI, energy, and geopolitics will be essential to navigating the next chapter of global transformation.