A single ChatGPT query consumes between 10 and 30 times more energy than a basic Google search. As artificial intelligence expands its reach and firms are integrating it into their technologies, its climate impacts are being felt. The data centres being built to power artificial intelligence search engines are placing pressure on local communities, driving up electricity costs, and straining water supplies.
As of March 2025, the U.S. has over 5,000 data centres. U.S. data centre electricity consumption is expected to account for 9% of the country’s electricity consumption by 2030, up from 4% in 2022, according to the International Energy Agency and EPRI. Major tech companies in the U.S. often opt to build their data centres in smaller towns near major cities; these companies receive tax incentives and benefit from relatively low electricity costs.
In Europe, data centre development is also expanding in key regions but is still struggling to meet demand. Major cities including Dublin, London, and Amsterdam have data centre investments, but firms are looking to expand into smaller cities as well. According to McKinsey, “Europe’s data centre power consumption is expected to almost triple from about 62 terawatt-hours (TWh) today to more than 150 TWh by the end of the decade.” This figure excludes cryptocurrency energy consumption, which is also significant.
Why does AI require so much power?
Data centres existed before artificial intelligence. They are used to store and process the data that the internet runs on, and they house the data from “the cloud”. Generative artificial intelligence (generative AI) must process far more data to produce answers and generate full reports. This demand increases as more workers use the technology to streamline their workflows. As AI’s processing capabilities increase, it requires more electricity and, in turn, more water to prevent the systems from overheating. Most data centres rely on evaporative cooling systems to keep the systems cool, a need greater in the summer when water consumption increases in general and droughts occur. Freshwater is necessary for the task of cooling the servers and facilities to prevent sediment and salt buildup during the cooling process.
Tech companies’ water usage has already vastly expanded as they invest more heavily in AI. Between 2021 and 2022, Microsoft’s global water use increased by more than 30%, while Google’s increased by 21%. According to a study entitled Making AI Less “Thirsty”: Uncovering and Addressing the Secret Water Footprint of AI Models, researchers estimate that “water withdrawal of global AI is projected to reach 4.2 – 6.6 billion cubic meters in 2027”, which is half of the total annual water withdrawal of the United Kingdom. The study also states that “0.38 – 0.60 billion cubic meters of water will be evaporated and considered ‘consumption” due to the global AI demand in 2027.”
In July 2023, residents in Uruguay protested a proposed Google data centre expected to consume vast quantities of water amid an extreme national drought. Google began construction on the project late last year, assuring residents of its commitment to sustainability. In the U.S. state of Georgia, since the construction of a massive Meta data centre, residents who live nearby have experienced extremely low water pressure in their homes, found sediment in their water supplies, and seen their electricity bills almost double. Some local politicians want the companies to be responsible for generating their own power, rather than putting strain on the main grid, while others see the investment as a boon for the greater community, as taxes paid by the tech giants increase the town’s revenue.
Sustainability vs. Expansion
Despite technology companies aiming to reduce the amount of energy their equipment requires to run by innovating towards more efficient models, the expansion of the technology and people’s adoption of it for work and personal use may not keep pace with companies’ or governments’ sustainability goals. European countries tend to have more regulation and less physical space to house the data centres than the U.S., but the competition for the technologies of the future is not slowing down and may incentivise rapid expansion over smart growth of an industry which requires vast resources to function. The AI revolution is not slowing down, and neither is climate change; unfortunately, the former seems bound to make the latter worse.
Sarah Sakeena Marshall, Data Assistant
