Renewables Alone Cannot Power AI Infrastructure

The Transformation of U.S. Electricity Demand

For decades, U.S. electricity demand grew at a predictable, modest pace. Utilities could plan around gradual increases driven by population growth and economic activity. But that era is over. A dramatic shift is underway, one that could define the next decade of energy infrastructure—and investment opportunity.

At the heart of this transformation lies the exponential growth of artificial intelligence (AI) and its insatiable appetite for electricity. The rise of AI-powered tools—especially those using large language models and machine learning—has created unprecedented demand for computing power. These complex models require massive amounts of electricity not just during training, but continuously during deployment (or inference) as well.

The Impact of AI and Electrification

This isn’t just about a few extra servers. AI data centers are power-hungry operations, often requiring the energy equivalent of a small city. Estimates suggest that U.S. data centers could consume more electricity by 2030 than entire countries such as Japan or Turkey. AI isn’t the only driver. The ongoing electrification of vehicles, heating systems, and industrial processes is also pushing grid demand to levels not seen in a generation. Together, these trends are accelerating what many experts are calling a new era for American energy.

Challenges and Solutions

From Texas to Northern Virginia, electric utilities are scrambling to adapt. Regions that already host dense clusters of data centers—like Virginia’s so-called “Data Center Alley”—are hitting capacity limits. Utilities are being forced to rethink everything: where and how they build generation, how they plan grid expansions, and how to regulate and forecast demand. While renewables like wind and solar will play an important role in the energy future, they alone cannot power a 24/7 AI infrastructure.

That’s why natural gas and nuclear are regaining prominence in grid planning. AI’s power requirements could also increase support for energy storage solutions and grid-balancing technologies. Several utilities have fast-tracked proposals for new natural gas peaker plants. Others are evaluating small modular nuclear reactors (SMRs) as potential solutions for delivering steady, low-carbon baseload power.

Investment Opportunities and Risks

For investors, this shift represents a seismic change. Grid operators, gas suppliers, and infrastructure developers stand to benefit as capital spending surges. Electricity producers may see dramatic increases in earnings. However, the investment story isn’t without risk. Regulatory hurdles, permitting challenges, and community opposition to new infrastructure could delay projects.

It’s crucial to understand the difference between regulated and unregulated utilities. Regulated utilities offer predictable dividends, while unregulated utilities are more exposed to market volatility. In this new environment, unregulated utilities may be the biggest winners, capturing rising electricity prices without regulatory constraints.

The Future of AI-Driven Demand

Forecasts indicate steeply upward demand growth, with data centers expected to double or triple their footprint by the end of the decade. State regulators and policymakers are grappling with how to accommodate this surge in demand while ensuring grid reliability and environmental concerns are addressed. Resilience is becoming a key focus, particularly in the face of extreme weather events and cyber threats.

Artificial intelligence's impact on the physical world, particularly the energy grid, is profound. Understanding the role of electricity in enabling AI innovation is crucial for investors looking to position their portfolios for the future.