- Global electricity consumption is projected to rise sharply over the next decade due to the rapid growth of AI. Data centers supporting AI systems are expected to consume more power than Japan by 2030.
- In the U.S., data centers are projected to account for nearly half of the country’s electricity demand growth through 2030, potentially surpassing the combined energy consumption of processing aluminum, steel, cement and chemicals.
- Emissions from global data centers could rise significantly, but the IEA believes the broader deployment of AI technologies could lead to emissions reductions that outweigh the increase from data centers.
Full Story
Global electricity consumption is on track to rise sharply over the next decade, largely due to the rapid growth of artificial intelligence (AI), according to a new report by the International Energy Agency (IEA).
“AI is coming. In many sectors, it is already here. This has major consequences for the global energy sector,” the IEA report states. “There is no AI without energy–specifically electricity. At the same time, AI has the potential to transform the sector’s future.”
How much energy is AI projected to use?
The IEA projects that by 2030, electricity demand from the world’s data centers, which support AI systems and cloud computing infrastructure, will more than double. The report states data centers will consume more power than the entire nation of Japan. Meanwhile, those data centers optimized for AI will see their energy use quadruple by the end of the decade.
A typical AI-focused data center can use as much power as 100,000 homes, and some of the largest currently under construction may consume up to 20 times more than that. Since 2017, energy needs from data centers have grown four times faster than overall global electricity demand and now account for approximately 1.5% of worldwide power usage.
How will data centers impact US energy demand?
In the United States, data centers are projected to account for nearly half of the country’s electricity demand growth through 2030. By then, the electricity used for American AI-related data processing needs is expected to surpass the combined energy consumption of processing the nation’s aluminum, steel, cement and chemicals.
The geographic concentration of data centers also raises concerns. Developers are building about 50% of new U.S. facilities in already-established regional hubs, potentially leading to local energy bottlenecks and infrastructure challenges.
How will data centers impact the environment?
The IEA estimates that emissions from global data centers could rise from 180 million tons of CO₂ today to 300 million tons by 2035 — comparable to the total annual greenhouse gases produced by nearly 70 million gasoline-powered cars.
What happens next?
Despite these figures, the IEA notes concerns over AI accelerating climate change seem to be “overstated.” The agency believes the broader deployment of AI technologies could lead to “emissions reductions that are far larger” than those created from data centers.
