The race to power artificial intelligence is reshaping the U.S. electricity market, driving up competition for existing coal plants and pushing utilities into a bidding war that threatens to raise costs and complicate climate goals. As data centers multiply to handle AI workloads, the sudden surge in demand is reviving interest in coal-fired generation that many had expected to fade.
Coal plants become a hot commodity
Several aging coal plants that were slated for retirement are now attracting multiple buyers. The bidding war reflects a broader scramble for reliable baseload power that can run around the clock — something renewables like wind and solar can't yet guarantee. Utilities and tech companies alike are competing to secure long-term power purchase agreements, driving up prices for electricity from existing coal units.
This renewed demand comes as a surprise to many in the industry. Just a few years ago, coal was seen as a dying asset, squeezed by cheap natural gas and stricter environmental regulations. Now, the AI boom is giving some of these plants a second life, at least temporarily.
AI's insatiable appetite for electricity
Training large language models and running inference at scale requires enormous amounts of electricity. A single AI data center can consume as much power as a small city. That demand is growing fast, and it's colliding with a power grid that wasn't built for this kind of load.
Utilities are finding that they can't bring new renewable projects online fast enough to meet the immediate need. Natural gas plants are running harder, but even that isn't enough. So they're turning to coal — the most carbon-intensive source of power — to fill the gap.
Higher costs and tougher climate math
The bidding war is pushing up wholesale electricity prices, which will eventually hit consumers and businesses. At the same time, it's making it harder for utilities to meet their own sustainability targets. Many have pledged to cut emissions, but the AI-driven demand spike is forcing them to delay coal plant retirements or run them more often.
This tension between growth and decarbonization is playing out in real time. Some regulators are questioning whether new data centers should be required to pair their load with new clean energy, rather than relying on existing fossil plants. But so far, no clear policy has emerged.
The question now is how long the coal revival will last. If AI demand continues to climb, the pressure on the grid — and on climate goals — will only intensify. The next few years will show whether the market can find a way to power the AI revolution without derailing the energy transition.
