Hvor har vi vores viden om datacentre fra?
Datacentrenes effekt på vand- og energiforbrug er veldokumenteret, og der er megen forskning der peger utvetydigt på at ressourceforbruget i forbindelse med datacentre er i modstrid med påstande om at AI bidrager til den grønne omstilling. Udviklingen de senere år viser, at det omvendte er tilfældet: vores brug af AI, og de datacentre som er nødvendige for at drive den, accellererer vores energiforbrug, ikke den grønne omstilling. Læs fx denne artikel i Altinget :
Frem mod 2050 kommer Danmarks samlede elforbrug til at vokse sig mellem seks og syv gange større. Det anslår Energistyrelsen i dens opdaterede analyseforudsætninger til Energinet, hvor forbruget i 2050 vurderes at være på over 200 TWh.
Kilderne nedenfor dokumenterer eller diskuterer energiforbruget forbundet med datacentre. Datacenterkritisk Arbejdsgruppe er ikke interesserede i at overdrive negative historier eller overdramatisere, men i at give så retvisende et billede som muligt. Vi har ikke alle læst alle artikler i deres helhed, så skriv gerne til os, hvis nogle af kilderne er problematiske - og gerne hvis vi har overset vigtige ressourcer.
Ærindet med at formidle ressourcerne skal ikke forstås som et principielt angreb på digitalisering (eller fremskridt), fx peger forskere som Christoph Becker og Neil Selwyn på mange interessante alternative og bæredygtige tilgange til at udvikle og bruge digitale værktøjer.
Nogle af ressourcerne kommer fra en kurateret liste fra Critical Data Center Studies. En anden vigtig kilde har været "AI Skeptics Reading Group" og deres artikelsamling.
I afsnittene nedenfor har vi sorteret vigtige og opsigtsvækkende artikler i tre kategorier, hhv. onlineressourcer, forskningsartikler og journalistiske artikler.
Online ressourcer om datacentre
Datacente.rs: Interaktivt verdenskort over datacentre, her angives det at der er 81 datacentre i Danmark.
Launched in 2007, Data Center Map was the first research tool of its kind. We operate a global data center directory, mapping data center locations worldwide. Our intention is to make it easier for buyers, sellers, investors, regulators and other professionals working with the industry to gain insights into the markets of their interest.
As data center infrastructure rapidly expands across the United States, FracTracker’s National Data Centers Tracker reveals a growing threat: energy-intensive facilities operating with limited regulatory oversight, driving up electricity demand, pollution, and environmental injustice at the expense of ratepayers and frontline communities.
AI incidents are on the rise, yet current databases struggle with inconsistent structure, limiting their utility for policymaking. The AI Incident Tracker project addresses this by creating a tool to classify AI incidents based on risks and harm severity. Using a Large Language Model (LLM), the tool processes raw reports from the AI Incident Database and categorizes them using established frameworks, such as the MIT Risk Repository and a harm severity rating system based on CSET’s AI Harm Taxonomy.
Journalistiske artikler
Det statslige Energinet er under et umuligt pres for at udbygge den grønne energiinfrastruktur i takt med behovet. Nu kommer de store datacentre og kræver masser af strøm. Vi må have en diskussion om, hvem der skal have den grønne energi, siger borgere.
We spoke to two dozen experts measuring AI’s energy demands, evaluated different AI models and prompts, pored over hundreds of pages of projections and reports, and questioned top AI model makers about their plans. Ultimately, we found that the common understanding of AI’s energy consumption is full of holes.
Kilde: James O’Donnellarchive, and Casey Crownhartarchive. “We Did the Math on AI’s Energy Footprint. Here’s the Story You Haven’t Heard.” MIT Technology Review, . Accessed 13 Nov. 2025.
The Ecological Cost of AI Is Much Higher Than You Think As the microchips behind artificial intelligence grow in complexity, each generation requires more energy, minerals and water than the last, driving a ruinous cycle with no end in sight.
Kilde: https://www.truthdig.com/articles/the-ecological-cost-of-ai-is-much-higher-than-you-think/
A data center, which can use as much electricity as Philadelphia, is the new American factory, creating the future and propping up the economy. How long can this last?
Kilde: “Inside the Data Centers That Train A.I. and Drain the Electrical Grid….” Archive.Is, 30 Oct. 2025, https://archive.is/Guezl.
Ratepayers across the U.S. are facing rising electric bills — a trend that could be turbocharged by Wall Street’s growing effort to capture the electric utilities we all depend on.
Kilde: Seidman, Derek. “Data Centers Devour Electricity. Private Equity Is Buying Utilities to Cash In.” Truthout, 11 Nov. 2025, https://truthout.org/articles/data-centers-devour-electricity-private-equity-is-buying-utilities-to-cash-in/.
The AI investment boom driving roughly half of United States GDP growth in 2025 rests on shaky foundations, demanding extensive reorganization of public and natural resources to sustain an investment in data centers that underlies what some experts regard as an unsustainable AI bubble.
Kilde: Michigan Offers Handouts for Data Centers Promising Jobs. Will Those Jobs Come? | TechPolicy.Press. https://www.techpolicy.press/michigan-offers-handouts-for-data-centers-promising-jobs-will-those-jobs-come/. Accessed 12 Dec. 2025.
AI created as much carbon pollution this year as New York City and guzzled up as much H20 as people consume globally in water bottles, according to new estimates. The study paints what’s likely a pretty conservative picture of AI’s environmental impact since it’s based on the relatively limited amount of data that’s currently available to the public.
Kilde: Calma, Justine. “AI’s Water and Electricity Use Soars in 2025.” The Verge, 17 Dec. 2025, https://www.theverge.com/news/845831/ai-chips-data-center-power-water.
Forskningsartikler om datacentres energiforbrug
Artificial intelligence (AI) systems are rapidly becoming the key growth driver of global data center electricity consumption. Despite AI system power demand approaching that of a country the size of the United Kingdom, the environmental impacts of this growth remain unclear. Most assessments focus on the cost of interacting with specific AI models but do not provide a more holistic overview. Such estimates are complicated by the fact that data center operators do not publicly disclose the required inputs. Reports that attempt to address the global environmental impact of AI hardware typically rely on proprietary analyst data, limiting validation in the public domain.
Kilde: Vries-Gao, Alex de. “The Carbon and Water Footprints of Data Centers and What This Could Mean for Artificial Intelligence.” Patterns, vol. 0, no. 0, Dec. 2025. www.cell.com, https://doi.org/10.1016/j.patter.2025.101430.
Recent research by Alex de Vries-Gao shows that, by 2025, AI systems could have a carbon footprint comparable to that of a global city such as New York and consume as much water as all bottled water drunk worldwide in a year. Because tech companies withhold crucial data, the true environmental impact of AI largely remains out of sight.
Kilde: “AI’s Hidden Carbon and Water Footprint.” Vrije Universiteit Amsterdam, https://vu.nl/en/news/2025/ai-s-hidden-carbon-and-water-footprint. Accessed 19 Dec. 2025.
The rapidly increasing demand for generative artificial intelligence (AI) models requires extensive server installation with sustainability implications in terms of the compound energy–water–climate impacts. Here we show that the deployment of AI servers across the United States could generate an annual water footprint ranging from 731 to 1,125 million m3 and additional annual carbon emissions from 24 to 44 Mt CO2-equivalent between 2024 and 2030, depending on the scale of expansion.
Kilde: Xiao, Tianqi, et al. “Environmental Impact and Net-Zero Pathways for Sustainable Artificial Intelligence Servers in the USA.” Nature Sustainability, Nov. 2025, pp. 1–13. www.nature.com, https://doi.org/10.1038/s41893-025-01681-y
Det mangler vi viden om
Et tema, som det er meget svært at skaffe viden om, er hvad datacentre egentlig bliver brugt til. Er stigningen i energiforbruget legitimeret af de smartere, og kritiske velfærdsydelser som AI leverer? Eller bliver datacentre mest af alt brugt til streaming, porno, gaming, gambling og andre typer indhold som ikke har samfundsnytte og ikke bidrager til at levere løsninger til klimakrisen? Hvor er den viden? Eller er den principiel utilgængelig pga. lovgivning? Har du viden om hvad datacentre bliver brugt til, og proportionerne mellem need-to-have og nice-to-have på dem, så skriv gerne til os!
Opdateret 7.3.2026