AI
Farm animals use far more water, land and climate than AI
By WhyStopHarmingAnimals.com
AI data centers have drawn a lot of criticism for their water and power use. Measured against farm animals, their footprint is small[2][5]. Breeding and feeding tens of billions of animals uses about 200 times as much freshwater as AI data centers are projected to use[2], produces about 40 times the climate impact of every data center on Earth[5], and takes up land on the scale of continents.
Water
Wisser and colleagues estimated in 2024, in the journal Water, that animal agri-food systems consume about 228 cubic kilometres of freshwater a year[1][2], and that growing feed accounts for more than 90% of it[1]. Morgan Stanley’s analysis from September 2025, as summarized by FoodFacts, projected water use linked to AI data centers at about 1,068 billion litres a year by 2028, roughly 1 cubic kilometre[2]. 228 divided by 1.07 is about 213. The AI figure is a projection for 2028 and the farm figure is a current estimate, so the two don’t line up exactly, but a gap of about 200 times[2] is far too big for that to matter.
The U.S. figures agree. Lawrence Berkeley National Laboratory’s 2024 report found U.S. data centers used about 66 billion litres of water on site in 2023[3], with a range of about 150 to 275 billion litres projected for 2028[3]. The water used at power plants to make their electricity is larger than the on-site figure, and farm animals still use more even after that is added[3]. CBS News, drawing on figures from LBNL and the University of Twente, reported that Americans use more water to grow food, especially meat, than all U.S. data centers combined[3][4].
Climate
The International Energy Agency put the world’s data centers at about 415 terawatt-hours of electricity in 2024, about 1.5% of global electricity use, and the carbon dioxide from that power at about 180 million tonnes, on the order of 0.5% of the world’s CO₂[5]. Its base case for 2035 is about 300 million tonnes, still under 1.5% of energy-sector emissions. Carbon Brief’s charts of data-centre energy and emissions show the same scale[6].
FAO’s assessment of farm animals is in a different league: about 14.5% of all human-caused greenhouse gases[7], or about 7.1 billion tonnes of CO₂-equivalent[7], mostly from methane, manure and land cleared for feed[7]. 7.1 billion divided by 180 million is about 40[5][7]. The FAO figure covers all greenhouse gases and dates from its 2013 report, while the IEA figure is carbon dioxide from electricity in 2024. Other sources put the food system’s share in the same range: Poore and Nemecek put all food at about 26% of global greenhouse gases[8], and Xu and colleagues found animal foods and their feed account for about 57% of food-system emissions[9]. Data centers are a real and growing load on the grid. Breeding tens of billions of animals is a far bigger climate problem[5][7].
Land
About 44% of the world’s habitable land is farmed, about 48 million square kilometres, according to Our World in Data[8][10]. Grazing and feed for farm animals take about 80% of that farmland[8][10], an area on the order of the Americas. A worldwide vegan diet would shrink farmland from about 4.1 billion hectares to about 1 billion[8][10].
Industry tallies put the world at close to 10,000 data centers. Give every one of them 200 acres and the total comes to about 8,000 square kilometres, a fraction of one percent of the land used for farm animals. Some new data centers are being built on former fields, which is a real local planning question, but it does not compare with the global land taken by cows and their feed[10].
Local water stress
A single large data center can draw millions of gallons of water a day, and more than 40% of U.S. data centers sit in regions with high water stress. That matters to the towns nearby, and data centers should not be put on failing aquifers. It does not change the global ranking. Farm animals push a share of the world’s rivers past their freshwater limits, and most of that water goes to irrigating feed across whole river basins.
Killing
AI does not breed or kill animals for its output. Faunalytics, working from FAO data, counts about 87.9 billion land animals killed for food in 2024[15], about 241 million a day[15], and chickens alone were about 78.5 billion of them[15]. Add about 130 billion farmed fish a year and a wild catch estimated at 1.1 to 2.2 trillion fish, and the food system is the largest killing on the planet. Construction and power lines can harm wildlife, as any industry can, but that is a different kind of harm from breeding animals by the tens of billions in order to kill them[15].
Anyone worried about AI’s water, land or climate footprint has a far larger one within reach at every meal.
Sources
[1]Wisser et al., Water 2024
Water, 2024
[2]FoodFacts, AI vs livestock water
FoodFacts
[3]LBNL 2024 U.S. Data Center Energy Usage Report
Lawrence Berkeley National Laboratory, 2024
[4]CBS News, How much water does AI really use?
CBS News
[5]IEA, Energy and AI
IEA, 2025
[6]Carbon Brief, data-centre energy charts
Carbon Brief
[7]FAO, Tackling climate change through livestock
FAO, 2013
[8]Poore & Nemecek, Science 2018
Science, 2018
[9]Xu et al., Nature Food
Nature Food, 2021
[10]OWID, half of habitable land
Our World in Data
[11]Andy Masley, data-center land use
Andy Masley
[12]WSJ, AI data-center water
Wall Street Journal
[13]LBNL 2024 report
Lawrence Berkeley National Laboratory, 2024
[14]Alliance for the Great Lakes / Sentient
Sentient Media
[15]Faunalytics, global slaughter
Faunalytics / FAO
[16]OWID, animals slaughtered
Our World in Data
[17]Mood et al., farmed fish 2023
Animal Welfare, 2023