- Updated: February 20, 2026
- 7 min read
Trump Administration Reverses Mercury Standards as AI Data Centers Drive Power Surge
Coal‑fired power plants in the United States face renewed scrutiny as the Trump administration rolls back mercury standards while AI data centers surge in electricity demand.
Answer: The Trump administration has repealed the Mercury and Air Toxics Standards (MATS), effectively reviving aging coal plants at a time when AI‑driven data centers are pushing U.S. electricity demand to new highs, creating a perfect storm of increased environmental pollution, heightened health risks, and a setback for climate‑change mitigation efforts.
In a move that has alarmed environmental groups and industry analysts alike, the Trump administration announced the rollback of federal mercury and air‑toxics standards just as generative‑AI workloads are prompting a wave of new data‑center construction across the country. This policy reversal not only opens the door for older, dirtier coal plants to stay online but also amplifies the power‑grid strain caused by AI‑intensive computing.
Background on Trump Administration Policy Changes
The original Mercury and Air Toxics Standards were introduced by the EPA in 2012 under the Obama administration and later strengthened by the Biden administration in 2024. These rules capped emissions of mercury, arsenic, lead, and other hazardous pollutants from coal‑ and gas‑fired power plants.
In early 2026, the EPA, guided by the Trump administration’s deregulation agenda, issued a final rule that reverts the standards to their 2012 baseline, effectively eliminating the tighter limits imposed in 2024. The administration argues that the rollback will “save $78 million annually in compliance costs” and “accelerate the deployment of critical AI infrastructure.”
Key Timeline
- January 2026: EPA announces intent to revisit MATS.
- March 2026: Public comment period closes; industry groups lobby for leniency.
- May 2026: Final rule published, rolling back standards to 2012 levels.
- June 2026: Several coal plants receive extensions to remain operational.
Impact on Coal Plants and Mercury/Air‑Toxics Standards
Coal plants have historically been the largest source of mercury emissions in the United States, accounting for roughly 50 % of the national total. By weakening MATS, the administration removes the legal requirement for these facilities to install costly scrubbers and continuous emission monitoring systems.
What the Original Standards Covered
The 2024‑strengthened MATS limited emissions of:
- Mercury – a potent neurotoxin linked to developmental disorders.
- Arsenic – a carcinogen affecting skin, lungs, and bladder.
- Lead – harmful to the nervous system, especially in children.
- Acid gases (SO₂, NOₓ) – contributors to smog and acid rain.
Consequences of the Rollback
With the standards relaxed, the EPA estimates a potential increase of up to 1,200 tons of mercury emitted annually if all affected plants forego upgrades. This translates to higher concentrations of mercury in waterways, where it bio‑accumulates in fish and ultimately reaches human consumers.
Moreover, the rollback encourages utilities to postpone retirement schedules for coal units that were slated for de‑commissioning. The energy policy landscape is now tilted toward short‑term cost savings at the expense of long‑term public health.
Connection to Rising AI Data‑Center Power Demand
AI models such as GPT‑4, Claude‑3, and emerging multimodal systems require massive computational resources. A single high‑performance AI training run can consume as much electricity as a small town.
AI Compute Growth Statistics
According to industry reports, U.S. AI data‑center electricity consumption grew by 38 % in 2025**, driven by:
- Expansion of large‑scale language‑model training clusters.
- Deployment of inference servers for real‑time AI services.
- Increased demand for AI‑enhanced video generation and analytics.
Why Coal Plants Are Returning to Service
Data‑center operators often seek “reliable baseload power” to guarantee uptime. When renewable sources cannot meet peak demand, utilities turn to existing coal plants as a quick‑fix solution. The AI data centers themselves are lobbying for stable, low‑cost electricity, inadvertently encouraging the revival of older, higher‑emission generation assets.
For example, the Tennessee Valley Authority (TVA) announced in May 2026 that it would keep two coal units online beyond their planned retirement dates, citing “growing power demand from AI‑intensive workloads.” This decision mirrors a broader national trend where AI‑driven electricity needs are reshaping the generation mix.
Environmental and Health Implications
The combined effect of weakened mercury standards and surging AI‑related power demand creates a “double‑whammy” for public health and climate goals.
Mercury Toxicity and Human Health
Mercury exposure, even at low levels, can impair cognitive development in children and cause kidney damage in adults. The environmental impact of increased mercury deposition includes higher concentrations in fish, which disproportionately affect communities that rely on fishing for protein.
Air Pollution and Climate Change
Beyond mercury, coal plants emit sulfur dioxide (SO₂) and nitrogen oxides (NOₓ), precursors to fine particulate matter (PM₂.₅) and ground‑level ozone—both linked to respiratory illnesses and premature deaths. The rollback also stalls progress toward the United States’ 2030 carbon‑reduction targets, undermining commitments made under the Paris Agreement.
“The Trump administration is wiping out health protections critical for protecting children from toxins like mercury just to save the coal industry some money,” said Nicholas Morales, attorney at Earthjustice, in a press release following the EPA’s final rule.
How UBOS Is Addressing the AI‑Energy Intersection
While policy debates continue, technology providers are exploring ways to reduce the energy footprint of AI workloads. Enterprise AI platform by UBOS offers a suite of tools that enable developers to optimize model inference, schedule compute during off‑peak hours, and integrate renewable‑energy‑aware scheduling.
Startups building AI services can leverage the UBOS for startups program to access pre‑configured, energy‑efficient containers that automatically scale down idle resources. Small‑ and medium‑size businesses benefit from UBOS solutions for SMBs, which include built‑in monitoring dashboards that flag excessive power consumption.
Developers looking for rapid prototyping can use the Web app editor on UBOS together with the Workflow automation studio to orchestrate data‑pipeline tasks that run only when renewable energy is abundant, thereby lowering the carbon intensity of AI training jobs.
For teams focused on marketing AI, the AI marketing agents can generate campaign assets without requiring massive GPU clusters, thanks to efficient inference models that run on edge devices.
UBOS also provides a rich UBOS templates for quick start, including the “AI SEO Analyzer” and “AI Video Generator” templates, which demonstrate how to embed AI capabilities while keeping compute budgets in check.
Companies interested in voice‑enabled AI can explore the ElevenLabs AI voice integration or the OpenAI ChatGPT integration to add conversational interfaces without over‑provisioning server resources.
For developers who need to store and query large embeddings, the Chroma DB integration offers a low‑latency vector database that can be hosted on energy‑efficient cloud instances.
All of these capabilities are highlighted on the UBOS homepage, where you can also explore the About UBOS page to learn more about the company’s commitment to sustainable AI development.
What You Can Do Next
Policy makers, industry leaders, and tech professionals all have a role in mitigating the environmental fallout from this policy shift:
- Advocate for stricter state‑level mercury and air‑toxics regulations.
- Choose AI infrastructure providers that prioritize renewable energy and efficiency.
- Implement workload‑scheduling tools that align compute with low‑carbon periods.
- Support legislation that incentivizes retirement of high‑emission coal plants.
- Leverage platforms like UBOS to build AI solutions that are both powerful and eco‑friendly.
By staying informed and selecting responsible technology partners, you can help ensure that the AI revolution does not come at the expense of public health or the planet.
For a deeper dive into how AI is reshaping energy consumption and what you can do to future‑proof your operations, explore our UBOS partner program and discover resources tailored for sustainable AI development.
Stay updated with the latest UBOS tech news to see how the industry evolves in response to these critical policy changes.