America has the energy resources to power AI, but outdated permitting rules are preventing new nuclear energy from coming online fast enough.

Data centers need energy. The United States has plenty of nuclear and clean power. The real threat is the anemic permitting system failing to expand supply in time for the AI boom.

As artificial intelligence (AI) and data centers drive a rapid surge in electricity demand, regulatory bottlenecks embedded in federal and state permitting rules have become a binding constraint on growth. While laws such as the Standardizing Permitting and Expediting Economic Development Act (SPEED Act) acknowledge these delays, they tend to focus on accelerating the journey of new energy infrastructure through an overly complex system, rather than simplifying the system once and for all. The result is a growing gap between available technology and deployable capacity. It’s beginning to threaten US competitiveness in the AI era.

Permitting Is the Binding Constraint on AI-Era Energy Growth

Building new nuclear and clean energy generation in the United States requires navigating federal permitting processes, which are notoriously slow. At the heart of this blockage lies a cumulative regulatory framework, anchored in legislation such as the National Environmental Policy Act (NEPA), which governs environmental review across major federal energy projects. Projects must navigate multiple agencies, successive environmental reviews, and overlapping veto points, creating delays incompatible with the scale and urgency of the investment required to meet rising energy demand.

Rather than reforming this underlying framework, recent policy responses have focused on narrower, agency-specific fixes. The Accelerating Deployment of Versatile, Advanced Nuclear for Clean Energy Act (ADVANCE Act), for example, aims to streamline nuclear licensing procedures within the Nuclear Regulatory Commission (NRC). Those challenges in the nuclear sector are real. Obstacles include excessively long licensing timelines, high regulatory costs, and rules poorly adapted to new technologies. But the ADVANCE Act’s remit is limited and short-sighted. It confines itself to the NRC’s internal processes, leaving the overarching environmental framework, which continues to constrain the approval of new energy projects, practically untouched.

The SPEED Act represents an attempt to address these bottlenecks at the NEPA level. It introduces reforms that would limit the scope of environmental reviews, shorten judicial timelines, and incorporate Supreme Court precedents. And still, it preserves NEPA’s basic structure, leaving overlapping veto points in place. Instead of truly simplifying the system, it speeds up the existing one—an important step forward, but a much smaller one than the sector needs, given the urgency of AI energy needs.

Big Tech Is Locking In Existing Nuclear Power, Not Creating New Supply

As a result, private companies are increasingly competing for existing energy capacity, particularly nuclear power, securing supply through long-term contracts instead of investing in new generation. In January 2026, Meta announced long-term agreements with several nuclear-sector partners, which could enable up to 6.6 gigawatts of capacity by 2035, positioning Meta as one of the largest corporate buyers of nuclear energy in US history.

The company relies on existing nuclear plants through long-term agreements with established providers such as Vistra and Constellation Energy, securing immediate or near-term power via life extensions and incremental capacity upgrades. Meanwhile, it is investing in more advanced nuclear technologies developed by specialized private firms such as TerraPower and Oklo. These new investments are unlikely to make a meaningful energy contribution until the 2030s.

The technology sector is squeezing as much energy as it can out of existing infrastructure, but if AI grows as fast as many predict in the coming years, there will come a breaking point at which demand will outstrip the available supply. Other major technology companies, including Microsoft, Google, and Amazon, are pursuing similar strategies, competing for clean and reliable nuclear baseload power. In practice, even with substantial private investment, most energy that can be delivered quickly still comes from existing capacity. Building new energy infrastructure simply takes too long.

Without Permitting Reform, AI Leadership Is at Risk

This structural bottleneck carries economic and strategic consequences. Electricity demand in the United States is rising at the fastest pace in decades, driven by AI-related data centers, while supply expansion remains sluggish. Without new capacity, key regions face prolonged grid interconnection delays, increased short-term reliance on natural gas, higher electricity prices for consumers, and growing risks to US leadership in artificial intelligence, a sector that is becoming increasingly critical from a national security perspective.

In this context, nuclear energy assumes a central strategic role. It’s a clean source of power. Crucially, unlike some renewable sources of energy, it provides firm, continuous electricity, an essential requirement for data centers which demand high, round-the-clock energy loads.

The regulatory framework urgently needs structural reform, not more attempts to manage complex bureaucratic mazes. Reforms like the ADVANCE Act and the SPEED Act are positive signals, but they don’t go nearly far enough. We are fiddling while Rome burns.

Without a deeper simplification of the permitting system, better coordination between federal and state authorities, and a meaningful reduction in cumulative veto points, the United States risks undermining its competitiveness in artificial intelligence—not because of any lack of innovation or investment, but because of avoidable regulatory constraints.

About the author: Cláudia Nunes

Cláudia Nunes is a senior contributor with Young Voices, specializing in innovation, economic freedom, and regulatory reform. She is also a frequent contributor to the Foundation for Economic Education (FEE), where she publishes regular written analysis and produces video content on public policy and political economy. Her work focuses on the intersection of technology, markets, and public policy, with particular attention to European governance and transatlantic competitiveness. She has a background in communications and data analysis and is currently completing a certified training-of-trainers programme focused on policy education and public engagement. 

The post To Meet AI’s Energy Needs, America Needs Permitting Reform appeared first on The National Interest.