Trump’s AI Data Center Push Is Raising Pollution Fears, Former EPA Officials Warn

Updated September 12, 2026 6:23 pm

Former Environmental Protection Agency officials say the Trump administration is loosening pollution safeguards to accelerate AI data center construction, putting Americans’ health at greater risk. In a new report released this week, the group argues that deregulation, cuts to environmental enforcement, and support for fossil-fuel power are creating a dirtier foundation for the country’s AI buildout.

The warning matters because the electricity-hungry expansion of data centers is not only a tech story but also a public-health and energy-policy fight. The officials say the administration’s drive to make the United States the “AI capital of the world” is being paired with weaker air and water protections at exactly the moment when more power plants, more chip production, and more industrial pollution are coming online.

Why former EPA officials are sounding the alarm

The core concern is simple: the rapid expansion of AI infrastructure is increasing emissions, and the Trump administration is making it easier for those emissions to rise. The critique comes from the Environmental Protection Network, a nonprofit made up of former EPA staff members who argue that the agency has shifted away from its mission to protect people and the environment.

According to the group, federal policy changes since January 2025 have already made it more likely that new data centers will be powered by dirtier electricity and built under looser environmental oversight. The result, they say, is a wave of pollution that may affect people far beyond the neighborhoods where data centers themselves are located.

“You would think that an industry buildout of this magnitude, billed as a revolution, would mean more vigilance from EPA,” said Lynn Goldman, a pediatrician and former EPA assistant administrator. “Instead, EPA is doing just the opposite, thinking first about industry’s needs and weakening or reconsidering limits on toxic chemicals and harmful pollutants.”

What the report says

The report identifies 30 federal actions since January 2025 that the authors say increase health risks tied to data center pollution. Seventeen of those actions specifically mention AI or target data centers, according to the group.

Those actions include executive moves and regulatory proposals aimed at slowing renewable energy deployment, easing pollution limits, and exempting some existing factories and power plants from certain standards. The former officials argue that these changes make it easier to support the massive electricity demand created by AI systems without accounting for the health costs.

The organization says the EPA should be tracking the pollution consequences of the data center boom, making the information public, and enforcing standards that limit harm. Instead, its authors say, the agency is weakening oversight just as the industry is scaling up.

How AI data centers are changing the pollution picture

AI systems rely on large data centers packed with servers, cooling equipment, and power-intensive hardware. As the industry grows, so does demand for electricity, and that can translate into more gas generation, more industrial emissions, and more exposure to pollutants such as nitrogen oxides, particulate matter, and toxic byproducts from fossil-fuel production.

The former EPA officials say the health risk is not confined to communities adjacent to a data center campus. Data centers draw on regional power grids, and the pollution associated with those grids can travel across county and state lines. Chip manufacturing and related industrial processes can also release pollutants that affect broader populations.

Goldman said people do not need to live next to a data center to be affected because the emissions come from the broader infrastructure supporting it, not just the server buildings themselves.

That point is important because many public debates focus on noise, land use, or local electricity demand. The EPA alumni are pushing the discussion outward, arguing that the health footprint of AI is much larger than the footprint of a single facility.

What makes this different from past tech booms?

This buildout is different because it combines several environmentally intensive sectors at once: digital infrastructure, power generation, semiconductor manufacturing, and industrial chemicals. Each one has its own emissions profile, but together they can amplify regional pollution in ways that are difficult to reverse once construction begins.

The report says the current policy direction is especially risky because it arrives alongside reductions in staffing and funding at the EPA. That combination, the former officials argue, weakens the government’s ability to monitor pollution and enforce the standards that are still on the books.

What Trump’s AI agenda is changing

The Trump White House has framed AI infrastructure as a national priority. EPA Administrator Lee Zeldin has said the administration wants to make the United States the global center of AI development, and the agency has moved to roll back or reconsider multiple environmental rules in that context.

In July 2025, the administration released an AI Action Plan that called for “streamlining or reducing regulations” under major environmental laws, including the Clean Air Act, the Clean Water Act, and Superfund-related rules, to speed permitting for data center projects and chip factories.

Critics say that approach effectively treats environmental permitting as a barrier to be removed rather than a safeguard to be honored. They argue that the administration is using emergency-style authority and deregulatory tools to accelerate fossil-fuel-heavy infrastructure for the AI sector.

Tyson Slocum, energy program director at Public Citizen, previously said the plan was really about opening the door to major exemptions for data centers and fossil-fuel infrastructure.

Which policy changes matter most?

The report highlights three broad areas of concern:

• restrictions on renewable energy development, which can push grids toward more fossil-fuel generation;
• relaxation of pollution standards for power plants and factories, which can increase emissions from the facilities serving data centers;
• faster permitting for chip factories and other industrial projects that support AI, even when those projects may produce hazardous waste or toxic chemicals.

Together, those shifts may not just increase total emissions. They may also change where pollution is concentrated, who is exposed, and how much health burden communities will bear.

Why fossil fuels and chip production are back in focus

The data center boom is also reshaping the energy market. As companies race to meet AI demand, states and utilities are facing pressure to add power quickly, often through new gas generation or on-site turbines. In some cases, plants are being designed specifically to serve large technology campuses.

At the same time, demand for advanced chips has revived attention on manufacturing processes that can involve potent chemicals with long environmental lifetimes. The report says that some of these materials do not break down easily and can accumulate in soil, water, and human bodies.

That is why the former officials describe the issue as a systemwide pollution problem rather than a narrow siting dispute. The environmental costs can be spread over long distances and long time periods, making them harder for the public to see — and easier for policymakers to ignore.

How the public-health impact could add up

The report cites research from UC Riverside, Caltech, and the Rochester Institute of Technology estimating that pollution linked to AI could cause as many as 1,300 premature deaths and more than $20 billion in public-health costs by 2028. Those numbers are projections, not certainties, but the authors say they illustrate the scale of the risk if current policy trends continue.

The Environmental Protection Network argues that the toll could be even higher if the administration keeps weakening rules and limiting enforcement capacity. In its view, the country still has time to steer the expansion toward cleaner power and tighter oversight, but only if regulators act now.

How much of this is preventable?

A great deal of it is preventable, according to the report’s authors. They argue that the U.S. could reduce the health damage from AI-related infrastructure by restoring clean air and clean water protections, strengthening enforcement, and making pollution data more transparent.

Marc Boom, the group’s senior director of public affairs, said the issue is not whether AI will grow, but whether it grows in a cleaner or dirtier way. He argued that EPA should be measuring the pollution consequences of the buildout, identifying exposed communities, and publishing the results.

“A more polluting AI buildout is not inevitable,” Boom said. “It depends on the choices government makes now.”

The group also wants the administration to adopt a “Data Center Health Protection Pledge,” a symbolic commitment that would signal a stronger emphasis on public health. The idea is unlikely to become law on its own, but the authors say it would at least force the White House to acknowledge the risks more directly.

Why EPA’s role is central

EPA’s mission is to protect human health and the environment, and former staff members say that mission is being diluted by the administration’s pro-development agenda. That criticism is especially pointed because the agency is supposed to be the federal backstop when industrial growth threatens air and water quality.

Earlier in the Trump presidency, hundreds of current EPA employees sent an open letter to Zeldin saying they believed the agency under his leadership would not adequately protect communities from dangerous chemicals and unsafe drinking water. The new report builds on that concern, arguing that the agency is now failing to act as a watchdog during one of the largest infrastructure buildouts in years.

The dispute highlights a broader political question: whether the AI race should be treated as an environmental emergency in its own right. Supporters of fast-track development see regulatory relief as necessary to compete globally. Critics say that approach shifts the costs onto communities that may never use a single generative AI tool but will still breathe the polluted air that powers it.

What the administration says in response

The EPA rejects the accusation that it is sacrificing public health for technology growth. In an email response, agency spokesperson Cora Mandy said the administration believes it is returning regulations to the proper reading of the Clean Air Act after what it views as years of overreach.

Mandy said the EPA remains committed to protecting human health and the environment while also backing the goal of making the United States the AI capital of the world.

That defense reflects the administration’s broader message: environmental rules should not stand in the way of innovation, investment, and industrial expansion. But the former officials say that framing misses the real issue — the public health costs of building AI on a dirtier energy and manufacturing base.

Key facts at a glance

The bigger picture

The fight over data centers is becoming a proxy battle over the future of industrial policy in the AI era. On one side are companies and officials who want the United States to move faster, build more, and remove what they see as regulatory drag. On the other are health and environmental advocates warning that speed without safeguards will deepen existing pollution burdens.

The stakes are especially high because AI infrastructure is still in a buildout phase. Decisions made now about power sources, permitting standards, and enforcement could lock in emissions for years or decades. Once a gas plant, turbine network, or chip factory is built, those emissions do not disappear when the policy debate moves on.

That is why the former EPA officials are asking for action now, not later. They are trying to shift the conversation from innovation at any cost to innovation with accountability. Whether the administration changes course is another question entirely.

For now, the message from the report is blunt: the AI boom may be accelerating, but without stronger environmental guardrails, so are the pollution and health risks that come with it.

Timeline of the dispute

What happens next?

The immediate next step is political rather than technical: whether the White House, EPA, or Congress will respond to the warning with tighter oversight or keep prioritizing rapid expansion. The report’s authors are not expecting a dramatic reversal, but they are trying to keep public pressure on the issue.

That could mean more scrutiny of power plants, more debate over where AI infrastructure should be built, and more attention to the public-health cost of the race to dominate artificial intelligence. For now, former EPA officials are making the case that a cleaner AI future is still possible — but only if policymakers choose it.