Industrial resource consumption is not a new bone of contention, and progress has always had its downsides. The question is how long the positive aspects of progress will outweigh the negative ones, and how we can ensure that this remains the case.
There is always a shortage of water somewhere in the world
According to the Finnish Water Utilities Association FIWA (2021), only about 2.5% of the Earth’s water is fresh, and most of it is locked up in glaciers, meaning that only a fraction of fresh water is available for human use. In many countries, rainfall is also lower than consumption, and although water is a renewable resource and there is enough fresh water on Earth for everyone, it is difficult to distribute it evenly. The amount of water resources, water supply coverage, and water use in households, industry, and agriculture vary from country to country, and in 2021, it was estimated that by 2025, up to two-thirds of the world’s population will live in areas with intermittent water shortages. (FIWA, 2021) Industry consumes large amounts of water, and mining is one of the most water-intensive industries (Water Europe 2021). Climate change is exacerbating the water situation in many countries, and in the 2020s, the rapidly trending artificial intelligence technology has also become a topic of discussion in the water debate.
Thirsty AI
Data centers are huge warehouses full of networked servers that are used for remote data storage and processing, as well as by IT companies to train AI models such as ChatGPT. Data centers are an important part of the information network infrastructure, as they run banking systems, patient databases, and other important functions (Sallinen, 2023), and modern society cannot function without them. The rapid growth of AI applications requires enormous data center capacity, which is not without its problems: data center backup power supplies, servers, computers, cooling systems, systems that regulate conditions such as lighting and humidity, and various monitoring equipment consume a significant amount of electricity (Paukkeri, 2021). High electricity consumption generates heat, and since it is essential for data centers to maintain stable conditions for their equipment around the clock (Paukkeri, 2021), the excess heat must be dissipated. This is where water comes into the picture, as data centers are often cooled with water, and indirect water consumption is also caused by, for example, water used in electricity production. Data centers are often built inland, far from the coast, as salt water in the air increases the risk of corrosion of metal parts (Barratt et al. 2025). In addition, to ensure physical security, they must be located in areas that are not prone to earthquakes or other natural disasters (Paukkeri, 2021), and as a result of these factors, they are increasingly being built in areas that already suffer from water shortages.
Out of the frying pan and into the fire
Tech giant Amazon is planning three new data centers in the Aragon region of northern Spain, and the data centers have been granted permission to use an estimated 755,720 cubic meters of water per year (Barratt et al. 2025). This corresponds to the annual water consumption of more than 15,000 Europeans (Eurostat, 2025). Southern Europe has suffered from severe droughts, intense heat waves, and widespread forest fires in recent years, resulting in local food production already suffering from water shortages (EEA, 2018). On the other hand, data centers are desirable projects in many municipalities because of the jobs, potential tax revenues, and other benefits they bring. Many of the tech giants building data centers have followed the example of food & beverage companies such as Coca-Cola (2024) and announced plans to save water, with the aim of becoming “water positive” by a certain year. However, when it comes to water, simple approaches such as “consume water elsewhere, reduce consumption or increase production elsewhere” cannot be taken – unlike carbon dioxide, for example, which is a global problem, water scarcity is local and transporting water over long distances is problematic (Barratt et al. 2025). Saving water on the other side of the world does not help communities suffering from local water shortages. In countries such as Finland, where there is sufficient water – and the cold climate required for cooling – building data centers is not as problematic.
One conversation with ChatGPT consumes at least half a liter of water. How many conversations do you have per day, and how many of them are critical?
Technology is constantly evolving, and so are energy efficiency solutions. However, it can be argued that the growing consumption of artificial intelligence is problematic per se: the increasing use of AI for entertainment consumes huge amounts of electricity and water, which are of course away from somewhere else – such as agriculture or industry. According to YLE, one conversation with ChatGPT consumes at least half a liter of water (Siltanen, 2025). How many conversations do you have per day, and how many of them are critical? Critical voices have also emerged, questioning whether thinking skills will deteriorate if all brainwork is outsourced to AI applications. What do you think?
Critical raw materials are a hot topic, too
Water and electricity consumption have been in the headlines regularly since the early days of the AI boom, but AI and its opportunities and threats are far from black and white. One side effect of the growth of AI is a significant increase in the demand for copper. The growing popularity of AI and, in particular, the construction of massive and dense data centers is driving up demand for copper: BloombergNEF predicts an additional demand of around 400,000 tons per year and a potential price spike by 2028 (Mining.com, 2025). This, combined with energy transition projects and the mining industry’s slowness to keep up with demand, could lead to a significant copper shortage. Copper is included in the EU’s list of critical raw materials (CRM). The EU defines critical raw materials as materials that are of high economic importance and whose availability is highly risky, including many metals and minerals needed for the green transition, digitalization, and industrial value chains (European Commission, 2024). There thus need for resource efficiency.
More of an opportunity than a threat?
Despite the risks and threats associated with it, artificial intelligence is generally seen as an opportunity. According to an expert interviewed by YLE, artificial intelligence has a significant role to play in halting climate change, and there is no reason to despair (Siltanen, 2025). How AI will solve climate change is not yet entirely clear, but we can do more than just trust that someone else (such as AI) will solve the problems for us. For example, the mining industry has the opportunity to improve its water efficiency—for example, through innovations such as EPSE—and the industry has indeed woken up to sustainability requirements. Artificial intelligence can also help save water. As stated, the issue is not black and white, and there are no easy answers. At EPSE, we strive every day to solve challenges related to water consumption, the circulation of metals and water, and a sustainable future. Future technologies must respond to the demand they create, and EPSE is committed to doing just that: with our method, we can save water, reduce emissions, recover metals, simplify processes, and create sustainable, cumulative, and measurable impacts.
References:
Barratt, L., Gambarini, C., Witherspoon, A. & Uteuova, A. (2025) Revealed: Big tech’s new datacentres will take water from the world’s driest areas. The Guardian, 9.4.2025. Available: https://www.theguardian.com/environment/2025/apr/09/big-tech-datacentres-water
Coca-Cola Finland (2024) Vesiprojektit. Coca-Cola – Vesihallinto (Water Stewardship). Available: https://www.coca-cola.com/fi/fi/sustainability/water-stewardship/vesiprojektit
EEA (2018) Water use in Europe — Quantity and quality face big challenges, European Environment Agency. Available: https://www.eea.europa.eu/signals-archived/signals-2018-content-list/articles/water-use-in-europe-2014
European Commission (2024) Critical Raw Materials. Single Market, Internal Market, Industry, Entrepreneurship and SMEs. European Commission. Available: https://single-market-economy.ec.europa.eu/sectors/raw-materials/areas-specific-interest/critical-raw-materials/critical-raw-materials_en
Eurostat (2025) Water statistics – Statistics Explained, European Commission. Available: https://ec.europa.eu/eurostat/statistics-explained/index.php/Water_statistics
Mining.com (2025) AI boom pushes copper toward a supply crunch, BNEF warns. MINING.COM, 25.8.2025. Available: https://www.mining.com/video/ai-boom-pushes-copper-toward-a-supply-crunch-bnef-warns/
Paukkeri, T. (2021) Datakeskuksen eri jäähdytysvaihtoehdot ja niiden kustannusvertailu Opinnäytetyö. Tampereen ammattikorkeakoulu. Available: https://www.theseus.fi/bitstream/handle/10024/495617/Paukkeri_Toni.pdf?sequence=3
Sallinen, P. (2023) Hukkalämpö on kahden kauppa. Energiauutiset, 7.9.2023. Available: https://www.energiauutiset.fi/kategoriat/markkinat/hukkalampo-on-kahden-kauppa.html
Siltanen, M. (2025) Datakeskusten valtava vedenkulutus on järkyttänyt maailmalla – Suomessa tilanne on toinen. Yle, 11.8.2025. Available: https://yle.fi/a/74-20175822
FIWA (toim.) (2021) Globaalit vesivarat: opettajan opas. EDUWATER, CBC-yhteistyöhanke. Available: https://www.vesi.fi/wp-content/uploads/2021/11/eduwater-globaalitvesivarat-opettajanopas2021.pdf
Water Europe (2021) The new CDP tool “Water Watch” ranks the most water-intensive sectors – Which are they?. Available: https://watereurope.eu/the-new-cdp-tool-water-watch-ranks-the-most-water-intensive-sectors-which-are-they/
This article was written by
Anni Honkonen
EHSQ Manager
anni.honkonen(a)epse.fi
Follow EPSE in social media!
