The rapid expansion of artificial intelligence (AI) infrastructure is emerging as an unexpected challenge to climate mitigation efforts in the United States. New federal data indicate that the soaring electricity demand from hyperscale data centers has slowed the country’s transition away from coal-fired power generation, contributing to a rise in greenhouse gas emissions from the electricity sector despite continued growth in renewable energy.

According to the latest report by the U.S. Energy Information Administration (EIA), carbon dioxide emissions from the power sector increased by 4 percent in 2025, reversing a long-term trend of declining emissions. Overall U.S. greenhouse gas emissions rose by only 2 percent during the same period, making the electricity sector the largest contributor to the increase. The report attributes much of this rise to a 13 percent increase in coal-fired electricity generation, driven in part by rapidly expanding demand from AI-powered data centers.

The findings highlight a growing paradox in the global energy transition. While artificial intelligence is expected to support climate research, optimize energy systems, and accelerate technological innovation, the infrastructure underpinning this digital revolution is placing unprecedented pressure on electricity grids and increasing dependence on fossil fuels.

For decades, electricity demand in the United States remained relatively stable, allowing utilities to gradually retire aging coal-fired power plants while replacing them with natural gas, solar, and wind energy. However, the rapid proliferation of AI applications has fundamentally altered electricity demand projections. Industry estimates suggest that data centers could consume more than 10 percent of the nation’s electricity by 2030.

Unlike conventional commercial facilities, hyperscale data centers operate continuously and require an uninterrupted power supply to support cloud computing, machine learning, and high-performance computing systems. Their round-the-clock energy demand often exceeds the availability of variable renewable sources such as solar and wind, particularly during nighttime hours or periods of low renewable generation.

As a result, utilities have increasingly relied on existing coal-fired power plants to maintain grid stability and ensure reliable electricity supply. Although renewable energy generation continues to expand rapidly, it is being used primarily to meet new electricity demand rather than replacing fossil fuel generation. Consequently, coal-fired plants that were previously scheduled for retirement are remaining in operation longer than anticipated.

The continued operation of coal-fired power stations carries significant environmental consequences. Coal remains the most carbon-intensive source of electricity generation and is also a major emitter of sulphur dioxide, nitrogen oxides, particulate matter, mercury, and other pollutants that contribute to poor air quality, acid rain, ecosystem degradation, and adverse human health impacts.

The increase in coal generation is occurring despite remarkable progress in renewable energy deployment. Solar and wind power recorded substantial growth during the past year, with solar energy surpassing coal generation during several months of high sunlight availability. These achievements demonstrate the increasing competitiveness and affordability of renewable technologies. However, the pace of clean energy deployment has not kept up with the extraordinary growth in electricity demand created by AI infrastructure.

Economic factors have further complicated the energy transition. Rising domestic natural gas prices, influenced partly by expanding exports, have narrowed the cost advantage of gas-fired electricity over coal. In several regions, existing coal plants have become comparatively economical to operate, encouraging utilities to increase coal generation rather than accelerate plant retirements.

Policy decisions have also influenced the trajectory of the power sector. Federal measures aimed at maintaining grid reliability have delayed the closure of several coal-fired power plants. While these interventions have attracted considerable public attention, available electricity generation data suggest that the recent rise in coal use is primarily being driven by increasing demand rather than regulatory actions alone.

The strongest increase in coal-fired electricity generation has occurred in regions experiencing rapid expansion of AI infrastructure. States hosting large clusters of hyperscale data centers have recorded substantial increases in electricity demand, with neighbouring coal-producing states supplying additional power to support these growing digital hubs. This regional shift illustrates how AI-driven electricity consumption is reshaping energy markets across multiple states.

The implications extend well beyond electricity generation. The power sector accounts for approximately one-quarter of total U.S. greenhouse gas emissions and is widely regarded as the cornerstone of national decarbonization efforts. Achieving climate targets depends heavily on reducing emissions from electricity production because cleaner power enables the electrification of transport, buildings, and industry, thereby reducing emissions across the broader economy.

Over the past two decades, emissions from the U.S. electricity sector declined by nearly one-third as utilities replaced coal with cleaner energy sources. This progress has been central to national climate strategies and international commitments to reduce greenhouse gas emissions. The renewed reliance on coal threatens to slow this momentum at a time when scientists warn that rapid emission reductions are essential to limit global warming.

The latest data also underscore the importance of expanding renewable energy deployment alongside investments in grid modernization, long-duration battery storage, flexible transmission networks, and other forms of reliable, low-carbon electricity. Without sufficient clean energy capacity available around the clock, growing electricity demand from emerging technologies risks prolonging dependence on fossil fuels.

As artificial intelligence continues to transform economies worldwide, its environmental footprint is becoming an increasingly important policy consideration. Ensuring that digital innovation is powered by sustainable energy sources will be critical for balancing technological progress with climate objectives. The latest federal findings serve as a reminder that achieving a low-carbon future requires not only rapid advances in clean energy but also careful planning to ensure that rising electricity demand does not undermine decades of progress in reducing greenhouse gas emissions.