AI Boom Could Double Data Center Power Consumption by 2030, Warns IEA
Electricity usage by data centers is projected to increase over two-fold by 2030, primarily due to the growth of AI applications, which will pose fresh challenges for maintaining energy security and achieving CO2 emissions targets, according to a statement from the IEA on Thursday.
The International Energy Agency (IEA), in its inaugural report on the impact of artificial intelligence (AI) on energy issues, stated that AI has the potential to facilitate more efficient production and consumption of electricity.
In 2024, data centers accounted for approximately 1.5% of worldwide electricity usage; this figure has risen by an annual rate of 12% over the last half-decade. Processing vast amounts of stored data necessitates immense computational resources, particularly for generative artificial intelligence systems.
Collectively, the United States, Europe, and China now make up roughly 85 percent of global data center energy usage.
Big tech companies increasingly recognise their growing need for power. Google last year signed a deal to get electricity from small nuclear reactors to help power its part in the artificial intelligence race.
Microsoft plans to utilize energy from new reactors at Three Mile Island, the location of one of America’s most significant nuclear accidents during the 1979 meltdown. Last year, Amazon entered into an agreement as well, aiming to employ nuclear power for their data centers.
At the current rate, data centres will consume about three percent of global energy by 2030, the report said.
As per the IEA, the power usage of data centers is projected to hit around 945 terawatt-hours (TWh) by 2030.
"The reported figures exceed Japan's current overall electricity usage. The primary catalysts behind this increase include AI along with rising demands for various digital services," according to the report.
A single 100-megawatt data center can consume an amount of energy equivalent to what 100,000 homes would use, according to the report. However, it also noted that upcoming data centers currently being built might require as much electricity as two million houses do.
The advisory group for energy policies based in Paris stated that "artificial intelligence could revolutionize the energy industry over the next ten years, leading to an increase in global electricity consumption due to data centers, while simultaneously offering substantial chances to decrease expenses, boost competitiveness, and lower emissions."
Aiming to stay ahead of China in the realm of artificial intelligence, U.S. President Donald Trump has established a "National Council for Energy Dominance," which aims to increase electricity generation.
Currently, coal meets approximately 30 percent of the energy demands for data centers. However, renewables and natural gas are expected to boost their share due to lower costs and greater accessibility in major markets.
The expansion of data centers will lead to an inevitable rise in carbon emissions due to increased electricity usage, predicted to go up from 180 million tons of CO2 currently to approximately 300 million tons by 2035, according to the IEA. Despite this forecasted increase, these emissions would still constitute only a small fraction of the anticipated 41.6 billion tons of worldwide emissions for 2024.