Data centers are hitting a hard ceiling: many utilities cannot accommodate new projects until new power plants and transmission lines are built, a process that takes years longer than constructing the facilities themselves. As artificial intelligence deployment accelerates, the mismatch between data center growth and grid capacity is emerging as a critical constraint on U.S. technological competitiveness, according to a Congressional Research Service report released September 1, 2026.

The challenge is straightforward but urgent. Data centers require enormous quantities of electricity to operate, and the federal government has made their development a priority as a core component of AI infrastructure. Yet the electrical grid, built over decades with different demand patterns in mind, is struggling to keep pace. Some utilities are experiencing infrastructure constraints so severe that they have stopped accepting new data center connection requests entirely until additional grid infrastructure is built.

How Much Power Do Data Centers Actually Consume?

Understanding data center electricity use requires breaking down where the power actually goes. Roughly half of a data center's electricity demand comes directly from running the computer servers and equipment that process and store data. The other half is split between cooling systems, which account for approximately 40 percent of electricity consumption, and facility operations like lighting and security systems.

The scale of these facilities has grown dramatically. Modern "hyperscale" data centers contain at least 5,000 computer servers and occupy at least 10,000 square feet of physical space, though many are far larger, occupying a million square feet or more. These massive installations are designed to handle the computational demands of AI systems, which require enormous data storage and processing capacity.

Why Is the Grid Struggling to Keep Up?

The core problem is timing. Building a data center typically takes less time than building the power plants, transmission lines, and distribution infrastructure needed to supply it with electricity. This creates a bottleneck where utilities cannot approve new data center projects because the grid infrastructure to support them does not yet exist.

The anticipated pace of U.S. electricity demand growth now exceeds anything seen in the last 20 years, raising fundamental questions about whether existing regulatory processes and infrastructure planning are suited for this acceleration. Many observers in the technology industry and government officials, including the Trump Administration, view grid infrastructure constraints as a significant hurdle to U.S. competitiveness in emerging technologies like AI.

What Are Policymakers Doing About This Challenge?

Congress has taken notice. The 119th Congress has held hearings examining data center and grid-related issues, with particular focus on ensuring that data center expansion does not raise electricity costs or compromise grid reliability for other consumers, especially residential customers. Both federal and state governments are actively developing policies to address the infrastructure gap.

The debate extends beyond technical concerns. Opposition to new data center construction is growing at the local level, with communities questioning whether the buildout of new grid infrastructure to support these facilities might lead to higher electricity prices for existing residents and businesses. This tension between technological progress and local economic impact is shaping policy discussions across the country.

Steps to Understanding Data Center Power Requirements

• Server Operations: Approximately 50 percent of data center electricity powers the computer servers and equipment that handle data processing and storage for AI applications and other services.
• Cooling Infrastructure: Around 40 percent of electricity consumption goes to cooling systems that maintain optimal operating temperatures for IT equipment, preventing overheating and system failures.
• Facility Support: The remaining 10 percent powers lighting, security systems, power distribution equipment, and other facility management systems necessary for safe and efficient operation.

The electricity grid itself operates through three interconnected systems: generation facilities that produce power, transmission lines that move electricity across long distances, and distribution networks that deliver power to homes and businesses. In most of the United States, utilities hold monopoly franchises in their service territories, meaning a single company controls all three functions and must have its rates approved by state regulators.

Some regions have restructured their electricity markets to allow competition in power generation, though transmission and distribution remain primarily controlled by regulated monopoly utilities. This patchwork of regulatory approaches adds complexity to data center siting decisions, as different regions have different rules for how new power plants connect to the grid and how electricity prices are set.

The infrastructure challenge is not merely theoretical. As data center demand continues to surge, utilities across the country are reporting that they lack the capacity to serve new projects. This creates a cascading problem: without new grid infrastructure, data centers cannot be built; without data center projects, there is less economic incentive to build new power plants and transmission lines. Breaking this cycle will require coordinated action from federal regulators, state governments, utilities, and technology companies to streamline permitting processes and accelerate infrastructure development.