Natural Gas Data Centers Pose Health Risks in Ohio, Study Finds

The hidden cost of AI's energy appetite

As artificial intelligence infrastructure expands across the United States, a troubling pattern is emerging: tech companies are increasingly powering massive data centers with dedicated natural gas plants that bypass traditional grid connections. This "behind the meter" approach now threatens to account for 40 percent of all data center power by 2030, bringing serious health and environmental consequences to surrounding communities.

Meta's New Albany, Ohio facility exemplifies the scale of the issue. The company installed 400 megawatts of dedicated natural gas turbines — enough capacity to power more than 300,000 homes — to run its data center operations independently of the local grid.

Quantifying the health impact

A study commissioned by the Piedmont Environmental Council examined a single Virginia data center facility and found stark results. The VA2 data center in Loudoun County could generate between $53 million and $99 million in annual health-related damages through increased deaths and hospitalizations in nearby areas, according to the research first reported by Ohio Capital Journal.

The problem extends beyond individual facilities. A shortage of gas turbines designed for power plants has pushed tech companies toward less efficient alternatives, often supplemented with backup diesel generators that produce even higher emissions. States like Ohio are fast-tracking approval for these installations despite the documented risks.

Why it matters

The rush to power AI with fossil fuels creates a false choice between technological progress and public health. While 70 percent of Americans oppose data centers being built near them, the industry's infrastructure decisions today will shape air quality, healthcare costs, and climate impacts for decades. Communities hosting these facilities face increased pollution from natural gas extraction, pipeline construction, and methane leaks, while all consumers likely face higher natural gas prices driven by surging demand.

Renewable alternatives already exist

China plans to power its data centers with at least 80 percent clean energy by 2030, even as it doubles AI infrastructure energy consumption. This demonstrates that renewable-powered AI development is feasible at scale.

U.S. tech companies have proven they can build clean data centers when they choose to. Google is constructing a Minnesota facility powered by 1.6 gigawatts of solar and wind paired with 300 megawatts of battery storage. Amazon is developing a 1.2-gigawatt solar and battery installation in Oregon for similar purposes.

Virtual power plant networks — coordinated systems of residential solar panels, home batteries, and electric vehicle chargers — offer another path forward. Tech companies could fund distributed energy installations and purchase excess power from residents during peak demand periods.

The Ohio Capital Journal originally reported these findings, highlighting the urgent need for policy changes as data center construction accelerates nationwide.