How The AI Boom Is Forcing A Clean Energy Reckoning

Artificial intelligence is forcing utility operators and policymakers alike to collaborate to address a potential energy dilemma—how to handle the expected growth in electricity demand from data centers.

But there is a fundamental contradiction between what the White House wants and market demand: The Trump Administration would like traditional baseload fuels like coal, natural gas, and nuclear to take the lead. But the grid screams for speed. It’s about keeping prices down, which means we need to bring on solar and wind energy—the resources that make up most of the electrons now in the interconnection queue.

“We have an administration hostile to clean energy: when you have rising demand, and you are not allowing the supply to increase, prices are going to go up. And right now, the administration is restricting new supply—the most quickly available new energy. The natural gas supply chain is constrained. The interconnection queues are full of solar and energy storage,” says Sean Gallagher, senior vice president of policy for the Solar Energy Industries Association.

His comments came during a symposium at the Intersolar and Energy Storage North America conference in San Diego. Solar remains among the cheapest sources of power: unsubsidized utility-scale solar—from construction to operation to decommissioning, known as ‘levelized costs’—is often less expensive than fossil fuels, according to the Energy Information Administration.

We are effectively paying a data center tax because AI is hogging the grid, and the administration is blocking the fast-build renewables that could ease the pressure. Electricity bills have already been climbing, and the average family is paying more just to ensure the grid doesn’t crash during a heatwave. Imagine the added pressure once the data centers come fully online.

Analysts estimate that this capacity spike will add 10–20% to the average bill in the 13 states that make up the PJM Interconnection—a regional grid serving about 65 million people across the Mid-Atlantic, Midwest, and parts of the South.

Utilities are rushing to build to meet 90,000 megawatts of peak-load growth projected through 2030. But if big tech finds a more efficient microchip, customers will get stuck with the bill for those “stranded costs”—expenses tied to infrastructure built for demand that never materialized.

This is compounded by the administration’s adversarial stance toward green energy: On Day One of his second term, Donald Trump issued an executive order freezing onshore and offshore wind energy projects. To that end, the Department of the Interior has officially paused construction on five major offshore projects, citing them as a national security risk. Industry analysts estimate that the combined effort will put 6 gigawatts of utility-scale power in doubt.

Trump refers to wind and solar as a “scam.” He stated on Truth Social that states relying on “WINDMILLS and SOLAR” face “RECORD BREAKING INCREASES IN ELECTRICITY AND ENERGY COSTS,” calling it “THE SCAM OF THE CENTURY!” The rhetoric matters because it shapes federal permitting decisions, funding priorities, and investor confidence across the energy sector.

A Generation Gap In The Making

Trump has declared a national energy emergency, calling for fast-tracking large 24/7 power plants. Yet these facilities take 5 to 10 years to build. Meanwhile, data centers are coming online within 24 months. By dumping on renewables, the administration is effectively creating a “generation gap”—a dangerous mismatch between when new power is needed and when the administration’s preferred sources can actually deliver it.

Consider the PJM Capacity Auction: It’s basically a massive “insurance premium” customers pay to ensure the lights stay on during the busiest times of the year. Think of the grid as a stadium, which is rarely at full capacity. But during the Super Bowl—peak demand—every seat is taken. The “capacity market” is what we pay to keep those extra seats built and maintained, even when they’re mostly empty.

In a year-end auction that sent shockwaves through the industry, PJM’s capacity prices hit a record ceiling of $333.44 per megawatt-day—a staggering 12-fold increase from just two years ago. The region found itself short of more than 6,600 megawatts of available capacity. Before this crisis, that same ‘seat’ cost about $28. At the same time, data centers are demanding huge blocks of power, causing PJM to revise its peak demand forecast upward by more than 5,200 megawatts.

Utilities don’t eat these costs. They pass them through to homeowners and businesses. The alarm bells reached a fever pitch in December 2025, when the U.S. Senate Banking Committee launched an investigation into how Wall Street is ‘cashing in’ on the AI boom. In a series of letters to BlackRock and Blackstone, Senators Elizabeth Warren, Bernie Sanders, and Richard Blumenthal specifically targeted the ‘golden age’ of utility acquisitions, arguing that private equity’s control of the grid is turning the AI revolution into a wealth transfer from everyday ratepayers to institutional investors.

“Affordability is really the watchword of the day, and we are really hammering that home, including an engagement with the administration,” says JC Sandberg, chief policy officer of American Clean Power, at the solar event.

In the current political climate, the discussion has shifted from climate to affordability, reliability, and resilience. The recalibration is practical. When electricity demand rises, the most persuasive argument is for cost stability.

“Our goal is 100% clean energy. Period. But it is not what we are talking about right now. Rates are going up. The rate of increase is outpacing inflation, and state leaders in red and blue states are feeling the heat,” says Heather O’Neill, CEO of Advanced Energy United, during the discussion.

Is Restraining Renewables Wise?

Renewables are not a silver bullet—but they are the fastest arrow in the quiver. They still require storage and transmission expansion, both of which demand significant investment. System integration adds to the complexity.

However, the more relevant question in the current demand cycle is not which resource posts the lowest levelized cost on paper, but which portfolio can deliver reliable megawatts at scale, quickly enough to prevent price spikes and volatility. Here, green energy is the fastest to market. But regulatory uncertainty complicates that equation. Industry estimates suggest roughly 78 gigawatts of wind and solar projects have been affected by federal policy constraints.

For investors and utilities alike, uncertainty raises the cost of capital and slows deployment decisions—leaving critical infrastructure in limbo at a bad time.

Storage, particularly long-duration batteries, is increasingly framed not as a climate instrument but as a hedge against volatility. By shifting excess midday generation to evening demand peaks, battery storage reduces reliance on expensive natural gas peaker plants. The good news is that EPRI—the Electric Power Research Institute—predicts the cost of long-duration energy storage will drop by 47% by 2030, with batteries capable of delivering 8 hours of power—and potentially more.

“What’s driving that? Scaling new tech. We’re seeing these extreme weather events, and we need to build resiliency. We need storage for these bigger projects to happen,” Anna Siefken, director of policy and markets for the Long Duration Energy Storage Council, told the audience.

The Coal Question

The coal question remains politically charged but financially straightforward. The president has moved to eliminate the “Endangerment Finding”—the 2009 EPA determination that greenhouse gases endanger public health and welfare. Eliminating it would remove the legal foundation for most federal climate regulations. Coal releases roughly twice as much CO2 as natural gas when burned.

Much of the existing coal fleet is more than four decades old. Operations and maintenance costs rise with age. Environmental-compliance retrofits require capital outlays that must compete with new-build alternatives. Extending the life of old assets may preserve short-term reliability, but it comes at the cost of investing in modern, efficient assets. Put simply: keeping old coal plants alive costs money that could be used to build faster, cheaper alternatives.

In the end, the defining feature of this moment is not ideology. It is about load growth and time-to-market. Artificial intelligence, electrification trends, and industrial expansion are reshaping electricity demand curves. Capital will flow toward the technologies that can scale under those conditions. The unresolved question is whether policy will facilitate that scaling—or constrain it at precisely the time flexibility matters most.

SIDEBAR:

The Strategic Cost Of A Stalled Grid

As of February 28, 2026, the global energy landscape has shifted. With the launch of Operation Epic Fury—the large U.S.-Israeli air strikes against Iran—the ‘affordability’ of the American power grid has transitioned from a domestic economic issue to a frontline national security concern.

While Tehran vows retaliation and the threat of global cyber warfare looms, the U.S. is sitting on a massive, untapped defensive asset: nearly 78 gigawatts of renewable energy currently stuck in permitting never land. This isn’t just lost revenue; it is a ‘Strategic Gap.’ A decentralized, renewable-heavy grid is inherently more resilient to cyber intrusions—seen recently in Poland and Italy. By letting these projects sit in a queue, we aren’t just spending more on power; rather, we are deliberately keeping our energy defenses weak at the moment the world enters a new era of open conflict.