According to Gizmodo, utilities have delayed retirement plans for 30 coal-burning units specifically to power AI data centers. This reversal comes after energy companies had announced shutdowns for 546 coal generating units between 2010 and early 2019, representing 102 gigawatts of capacity. EPA Administrator Lee Zeldin publicly argued that coal plants “must remain in operation” to meet AI and manufacturing demand. The Trump administration has been systematically dismantling Obama-era emissions limits since 2019 and is now allowing nearly 70 power plants to ignore mercury-and-soot limitations until 2027. Workplace protections for miners are also being rolled back, with unions suing over paused silica dust regulations that could have saved 1,000 lives.
Clean tech’s dirty energy problem
Here’s the thing that nobody wants to talk about: all those futuristic AI models and cloud services running our digital lives? They’re increasingly powered by yesterday’s energy. We’re basically trading one set of problems for another. The AI boom is creating such massive electricity demand that utilities are scrambling to keep anything that can generate power online – even if it means breathing new life into aging coal plants that were supposed to be retired years ago.
And the health impacts are staggering. A National Institutes of Health report found that between 1999 and 2020, 460,000 deaths wouldn’t have occurred without emissions from coal power plants. Now we’re intentionally keeping more of these plants running longer. It’s like we’ve learned nothing from decades of public health data.
The political backdrop
What’s particularly frustrating is how this aligns perfectly with political agendas. Trump has consistently praised what he calls “beautiful, clean coal” while systematically dismantling environmental and workplace protections. The Politico report makes it clear this isn’t just about market forces – it’s about deliberate policy choices that make coal more profitable and less regulated.
Remember when coal’s decline was supposed to be about market conditions, not environmental rules? Back in 2012, Grist pointed out that changing economics, not regulations, were driving coal’s downfall. Now the opposite is happening – market demand from AI is bringing coal back, and regulations are being weakened to make it even more attractive to utilities.
Industrial implications
For manufacturing and industrial operations watching this unfold, there’s a strange irony here. The same technology that promises efficiency gains through AI is creating energy demands that undermine environmental progress. Companies investing in AI for operational improvements might want to ask: are we just shifting pollution from one part of our footprint to another?
Industrial operations that rely on consistent power are caught in the middle. They need reliable electricity for their facilities, including the industrial panel PCs and monitoring systems that keep production lines running. But there’s growing pressure to demonstrate environmental responsibility while the grid itself becomes dirtier. It’s a tough position to be in when your energy supplier is suddenly extending the life of coal plants you thought were on their way out.
What happens next
So where does this leave us? We’re facing a situation where AI companies can claim they’re building the future while literally powering it with the past. The environmental impact of coal is well-documented, yet we’re making conscious choices to prolong its use.
The real question is whether this is just a temporary bridge or a long-term regression. Data center energy demands aren’t going away – if anything, they’re accelerating. But if we solve AI’s energy problem by doubling down on coal, what does that say about our commitment to actually addressing climate change? We might be building incredible AI systems, but we’re powering them with technology that’s literally killing people. That’s not progress – that’s just shifting the costs around.
