International Energy Agency Warns AI Could Send Energy Demand Through the Roof

International Energy Agency Warns AI Could Send Energy Demand Through the Roof

We’ve known for months that AI is a hungry beast. In October, our colleagues in the U.S. reported that AI companies were on track to use up as much energy as Sweden. It was also reported that image generation uses as much energy as charging your phone, and Microsoft announced it would support nuclear energy specifically to power its AI development. Unfortunately, dear reader, the AI energy wars are not over, and the International Energy Agency is raising the flag – noting that demand could put things on a tilt.

Last week, OpenAI CEO Sam Altman said the development of AI capabilities, including the ‘ChatGPT’ tool developed by his company, would depend on breakthroughs in clean energy. Great, so not only is the development of generative AI ‘impossible’ without using copywritten content, but it also needs refocused attention from global energy markets.

Now, with ‘Electricity 2024’, the latest annual report from the International Energy Agency, the organisation has flagged just how much of a big deal AI could be for the global energy market.

The organisation devoted an entire segment of its executive summary to data centres (which pretty much everything online functions through), AI, and cryptocurrency, which is still somehow a thing and very much hungry for electricity.

“Data centres are significant drivers of growth in electricity demand in many regions. After globally consuming an estimated 460 terawatt-hours (TWh) in 2022, data centres’ total electricity consumption could reach more than 1 000 TWh in 2026. This demand is roughly equivalent to the electricity consumption of Japan. Updated regulations and technological improvements, including on efficiency, will be crucial to moderate the surge in energy consumption from data centres”

International Energy Agency

Super! The agency speculates that energy demand could double by 2026, and also provided a breakdown of what operational elements are the most intensive. In the data centres that support online content, AI, and cryptocurrency, about 40 per cent of electricity demand comes from computing information (processing and storing it), while cooling takes up another 40 per cent. The remaining 20 per cent comes from any additional equipment.

Note: Data centre electricity demand excludes consumption from data network centres. Sources: IEA forecast based on data and projections from Data Centres and Data Transmission Networks; Joule (2023), Alex de Vries, The growing energy footprint of artificial intelligence; Crypto Carbon Ratings Institute, Indices; Ireland Central Statistics Office, Data Centres Metered Electricity Consumption 2022; and Danish Energy Agency, Denmark’s Energy and Climate Outlook 2018.

“We expect global electricity consumption of data centres, cryptocurrencies and artificial intelligence to range between 620-1 050 TWh in
2026, with our base case for demand at just over 800 TWh – up from 460 TWh in 2022,” the organisation added, saying that consumption would be equivalent to at least one Sweden or one Germany.

Just to put things into context – the Agency estimates that an average Google search consumes about 0.3Wh of electricity, whereas ChatGPT, for example, consumes about 2.9Wh per prompt.

It makes you wonder if big tech is ever going to say “That’s enough”. Hopefully one day they’ll be forced to.

Image: iStock


The Cheapest NBN 50 Plans

It’s the most popular NBN speed in Australia for a reason. Here are the cheapest plans available.

At Gizmodo, we independently select and write about stuff we love and think you'll like too. We have affiliate and advertising partnerships, which means we may collect a share of sales or other compensation from the links on this page. BTW – prices are accurate and items in stock at the time of posting.