In the heart of America’s energy grid, a surprising reversal is unfolding as data centers, fueled by the insatiable appetite of artificial intelligence, turn to coal—a fuel many thought was on its way out. High natural gas prices and skyrocketing electricity demands from AI workloads are pushing operators toward this carbon-intensive option, according to a recent report from The Register. The publication highlights how utilities in regions like the Midwest and Appalachia are delaying coal plant retirements or even ramping up production to meet the surge, with data centers projected to consume up to 8% of U.S. electricity by 2030.

This shift comes amid broader industry pressures, where tech giants like Google and Microsoft are racing to expand their computing infrastructure. The irony is stark: companies pledging net-zero emissions are indirectly bolstering the dirtiest fossil fuel, as coal’s reliability trumps intermittent renewables in the short term.

The AI Boom’s Hidden Energy Cost

Analysts at Jefferies, as cited in The Register’s analysis, note that natural gas prices have spiked 50% in the past year due to export demands and supply constraints, making coal economically viable again. Data centers, which can require gigawatts of power equivalent to small cities, are negotiating direct deals with utilities to secure stable supplies, often at the expense of cleaner alternatives.

Environmental advocates warn this could undermine U.S. climate goals, with coal emissions potentially rising 10-15% in key states by 2026. Yet, for operators, the math is simple: AI training models demand constant, uninterruptible power that solar or wind can’t always guarantee without massive battery backups.

Utilities Adapt to Surging Demand

The Department of Energy’s latest report, referenced in related coverage, forecasts domestic data center energy use doubling by 2028, driven by generative AI. This has utilities like those in Virginia and Texas extending coal plant lifespans, with some facilities that were slated for shutdown now being retrofitted to serve hyperscale campuses.

Industry insiders point to specific examples, such as Rainbow Energy’s plans in North Dakota to power a 300MW data center directly from the state’s largest coal plant, as detailed in reports from Data Center Dynamics. Such colocation strategies minimize transmission losses but lock in fossil fuel dependence.

Balancing Innovation and Sustainability

Tech firms are not idle; investments in nuclear small modular reactors and geothermal energy are accelerating, but these solutions won’t scale until the 2030s. In the interim, coal fills the gap, with global data center electricity demand expected to double by 2026, per the International Energy Agency’s findings echoed in The Register.

Critics argue this is a short-sighted fix, potentially inflating energy costs and regulatory scrutiny. For instance, Duke Energy’s updated carbon plan gambles on coal for data center growth, citing Trump-era rollbacks on pollution rules, as reported by NC Newsline.

Looking Ahead: Policy and Tech Intersections

As the sector evolves, collaborations like Babcock & Wilcox’s agreements to convert coal plants to gas hybrids signal transitional paths, though full decarbonization remains elusive. The McKinsey report on AI’s power hunger underscores opportunities for investors in resilient infrastructure, but warns of grid strains without coordinated policy.

Ultimately, this coal resurgence tests the tech industry’s green commitments, forcing a reckoning between rapid innovation and environmental responsibility. With AI’s exponential growth showing no signs of slowing, stakeholders must navigate these tensions to forge a more sustainable path forward.