Environment

US Datacenter Surge Matches Denmark Annual Emissions

US Datacenter Surge Matches Denmark Annual Emissions
Table of Contents
  1. Massive Emissions and Energy Consumption
  2. Reliance on Fossil Fuels
  3. Local Opposition and Health Concerns

The rapid construction of large-scale datacenters across the United States has driven planet-heating carbon emissions roughly equal to Denmark’s annual output, according to new research. The study highlights the heavy energy demands of artificial intelligence growth and its reliance on fossil fuels.

Massive Emissions and Energy Consumption

During the study period, 403 “hyperscalers”—the largest datacenters primarily designed to support artificial intelligence growth—released approximately 25.7 million metric tons of carbon dioxide. Published in Nature Sustainability, the research shows this pollution level matches the entire annual carbon output of Denmark, which has a population of 6 million, and represents a two- to threefold increase compared to a previous 2018 estimate.

The facilities accounted for about 1.8% of total US electricity consumption, according to the research, which utilized federal government data to track emissions over a one-year period leading up to April 2025. Researchers noted that because datacenter expansion has accelerated significantly since then, these figures likely represent a conservative baseline.

Francesca Dominici, a biostatistics professor at Harvard University and co-author of the study led by Gianluca Guidi, stated that the emissions are substantial and require policymakers to address the expansion responsibly.

Reliance on Fossil Fuels

The study estimates that the hyperscale group consumed between 68 and 99 terawatt-hours of electricity over the analyzed year. To meet this vast energy demand, the facilities largely turned to carbon-intensive power sources rather than established renewable options like wind and solar.

More than half of the electricity used by these datacenters originated from fossil fuels. The AI boom is increasingly supported by new gas-fired power generation, which is sometimes constructed directly at the datacenter sites.

Gina McCarthy, a former climate adviser to the administration of Joe Biden, described the study as a critical step in evaluating the impact of datacenters, noting that she was unsurprised by the heavy reliance on fossil fuels despite the associated climate and health consequences.

Local Opposition and Health Concerns

The expansion has sparked increasing alarm among climate advocates amid worsening global heatwaves, wildfires, and other climate impacts. While developers argue that these facilities create jobs, raise local taxes, and use marginal amounts of resources compared to industries like agriculture, public backlash continues to grow across the US.

In Tennessee, a cluster of datacenters tied to Elon Musk’s AI ventures has faced protests and legal action over noise and air pollution, leading the company to purchase nearby residences. Meanwhile, plans for Stratos, one of the world’s largest planned datacenters located in drought-affected Utah, have drawn sharp criticism over massive water consumption and air pollution.

A separate study concluded that Stratos could generate health-related economic costs reaching $690 million annually if it relies on the dirtiest fossil fuel options, releasing up to 3,140 tons of particulate matter that could affect communities across Utah and neighboring states.

John Walke, director of federal clean air at the Natural Resources Defense Council, which commissioned the report, warned that the project threatens severe health damages for local residents and millions living downwind. Stratos developers have planned to use combined-cycle natural gas while also evaluating renewables and other alternatives amid political pressure to scale down the project, recently filing a new water application claiming reduced water demand compared to past agricultural use at the site.

You May Also Like

Leave a Comment

Your email address will not be published. Required fields are marked *