Skip to content
IFAI

Industry

AI and crypto are set to double data center energy use

Published Reach 42

The International Energy Agency just warned that booming demand from AI and crypto mining could double global data center electricity use by 2026. Reaching up to 1,050 TWh, that growth is roughly equivalent to adding Germany's entire electricity demand to the grid, with AI alone forecasted to consume ten times more power. Why it matters: Computing appetites are outpacing local infrastructure, making clean power buildouts more urgent than ever.

AI and crypto are set to double data center energy use
Image: theverge.com

Every time you prompt ChatGPT or stream a video, somewhere a massive warehouse of servers gets a little hotter.

According to a new annual electricity report from the International Energy Agency (IEA), global data center power consumption could double by 2026, largely supercharged by artificial intelligence and cryptocurrency.

How much juice is that? The IEA projects data center demand could hit 1,050 TWh. That jump is roughly the equivalent of plugging an entire new country into the grid—think Sweden at a modest pace, or Germany at the high end.

AI searches and server cooling are drinking the grid

Back in 2022, data centers, crypto, and AI already swallowed about 2% of worldwide electricity, roughly 460 TWh. Crypto mining alone burned 110 TWh, nearly a quarter of that slice.

Inside a typical data center, electricity goes to two main appetites: computing hardware eats about 40%, and cooling systems burn another 40% just keeping the room from melting down.

AI is now acting like an accelerant on those numbers. The IEA notes that if Google fully integrates AI into every search, its electricity use could multiply tenfold. By 2026, the entire AI sector could burn ten times the energy it consumed last year.

Crypto isn’t slowing down either, with electricity demand forecast to grow 40% by 2026. While Ethereum managed to wipe out over 99% of its power consumption by shifting its validation model, the Bitcoin network refused to follow, driving the vast majority of mining emissions.

Regional grids are hitting a wall

This power grab is already breaking local infrastructure.

The US houses about 33% of the world's 8,000 data centers and leads the world in Bitcoin mining. The IEA expects American data centers to jump from 4% of total US power demand in 2022 to 6% by 2026, driven further by 5G rollouts and cloud computing.

Ireland is an even starker case. Thanks to rock-bottom corporate taxes, 82 data centers already sucked up 17% of the nation's power in 2022. With 54 more approved or under construction, tech facilities could chew through nearly one-third of Ireland's total electricity by 2026.

In London, data center power demands are actively stalling new housing projects. Down in Texas, an influx of crypto mines is putting massive strain on an already fragile power grid.

While the IEA does forecast that renewables will outpace coal by 2025 to generate over a third of global power, green energy additions alone won't keep up with this kind of surge.

Why it matters

AI's invisible bottleneck isn't algorithms or even microchips anymore—it's raw wattage.

For developers and everyday users, cheap compute might not last forever. As electricity becomes the primary constraint, cloud providers will inevitably pass rising utility bills onto API pricing and software subscriptions. In the near future, the most valuable models won't just be the smartest, but the most energy-efficient per token.

For the tech industry at large, the competitive battleground is shifting straight toward power infrastructure. Tech giants won't just fight for GPUs; they will be bidding against cities for grid capacity, substations, and dedicated clean-energy contracts.

We think the smart move right now is doubling down on efficiency rather than bruteforcing scale. Ethereum proved that altering core architecture can slash power burn overnight. If AI builders and Bitcoin networks keep ignoring the electric meter, they are likely walking right into aggressive regulatory caps and local power bans.

Sources (1)

  1. [1]
    theverge.com(opens in a new tab)Originaltheverge.com

Best of AI, daily