Data center carbon emissions in the United States have surged dramatically with the rapid growth of artificial intelligence and digital infrastructure. From 2018 to 2024, emissions increased roughly 300%, reaching about 105 million metric tons of CO₂ annually — a level that now makes data centers a significant contributor to national greenhouse gas emissions. Their average carbon intensity (548 gCO₂e per kWh) is much worse than the national grid average, largely because many centers are powered by fossil fuels.
Key Facts
- Emissions have surged: U.S. data center emissions have tripled from 2018 to around 105 million metric tons per year.
- Carbon intensity is high: Data centers emit about 548 gCO₂e per kWh, ~48% higher than average U.S. electricity generation.
- Fossil fuel reliance: Over half of the electricity used by data centers comes from fossil‑fuel‑based power plants.
- Growing global impact: Data centers worldwide may account for 1–3% of global emissions, with AI causing rapid increases.
- AI contribution rising: AI workloads are driving much of this growth, and by 2030, data center emissions could be many times higher without intervention.
Frequently Asked Questions (FAQs)
- Q: Why are data center emissions increasing so fast? A: The rapid build‑out of data centers to support AI and cloud computing has dramatically increased electricity demand — much of it powered by fossil fuels — leading to higher CO₂ output.
- Q: How big is the climate impact?
A: In the U.S., data centers now generate roughly 105 million metric tons of CO₂ per year, more than 2% of national emissions, and much of their electricity comes from fossil fuel plants.
- Q: Could these emissions be reduced?
A: Yes — switching to cleaner power sources, increasing energy efficiency, and adopting renewable energy can lower the carbon footprint of data centers.
- Q: How does AI specifically affect emissions?
A: AI‑focused workloads demand more compute power and energy, accelerating data center growth and pushing emissions higher, with some estimates projecting substantial future increases.
Resources/ Sources
- Alex de Vries-Gao: The Carbon and Water Footprints of Data Centers and What This Could Mean for Artificial Intelligence. Explains how data center carbon and water footprints grow as AI workloads expand and why their emissions are significant.
- Midhat Tilawat: AI Data Center Statistics 2026: The $200 Billion Revolution in Global Infrastructure. Highlights that global data center emissions are significant and projected to grow with AI demand, potentially reaching hundreds of millions of tons annually.
- Technology Watch: Environmental Burden of United States Data Centers in the Artificial Intelligence Era. Reports U.S. data centers consumed 56% fossil fuel–based electricity and emitted ~105 million metric tons of CO₂, ~48% more carbon-intensive than average U.S. power.
- Bad Data Centers- Carbon Emissions