- Data centre electricity demand could reach 945 TWh by 2030 equal to Japan’s current national consumption.
- These regions account for 85% of global data centre energy use, with the U.S. leading growth.
- CO₂ emissions from data centres may hit 300 million tonnes by 2035 despite gains in efficiency.
The International Energy Agency (IEA) has said artificial intelligence (AI) will drive a significant increase in data centre electricity use. By 2030, global data centres could consume over 945 terawatt-hours (TWh), matching Japan’s current annual electricity use.
The IEA’s new report links the surge to rapid growth in generative AI, which requires massive computing power and heavy data processing. In 2024, data centres used around 415 TWh of electricity, accounting for 1.5% of global demand. That figure has climbed by 12% yearly for the past five years.
If this trend continues, data centres will consume nearly 3% of global electricity by 2030. A single 100-megawatt data centre consumes as much power as 100,000 homes. Some new facilities reach 2,000 megawatts, enough to power 2 million households.
Developers often build these mega-centres near cities, putting pressure on local power grids. The United States, China, and Europe consume 85% of data centre electricity. The U.S. leads growth, with data centres set to drive nearly half of the country’s additional electricity demand.
To respond, the U.S. government launched the National Energy Dominance Council. The council aims to boost electricity generation and secure an edge over China in AI and data infrastructure. The IEA notes that coal currently powers 30% of data centres. However, energy providers are increasingly shifting to natural gas and renewables, which offer lower costs and broader access.
Despite cleaner alternatives, emissions from data centres continue to rise. The IEA projects that carbon dioxide (CO₂) emissions will jump from 180 million tonnes today to 300 million tonnes by 2035. However, this would still make up less than 1.5% of the total emissions from the global energy sector.
Some AI systems deliver efficiency gains, which help offset emissions. But the IEA urges caution. The agency says AI cannot singlehandedly solve the world’s energy problems. Generative AI brings benefits, but its high energy demands could cancel them out.
“No technology, including AI, can solve today’s energy challenges alone,” the report states. The IEA’s 302-page report marks its first complete analysis of AI’s impact on the energy sector.
Fatih Birol, the IEA’s Executive Director, called for urgent action. He urged governments to develop strong policies for AI development and energy use. “We need smart planning to manage AI’s impact and protect global climate goals,” he said.
The report encourages world leaders to balance digital growth with sustainable energy strategies. It highlights AI’s risks and potential benefits in the clean energy transition.
As AI expands, energy planners must act fast. Rising energy use could worsen emissions and strain electricity systems worldwide without strong policies.