A world of knowledge explored

READING
ID: 820XCE
File Data
CAT:Environmental Science
DATE:February 28, 2026
Metrics
WORDS:861
EST:5 MIN
Transmission_Start
February 28, 2026

Data Centers Threaten Climate and Water Resources

Target_Sector:Environmental Science

In a February 2026 letter to MPs, UK Energy Secretary Ed Miliband admitted something most politicians avoid: he doesn't know if the country's data centre boom will derail its climate goals. The impact on net zero efforts, he wrote, is "inherently uncertain." This wasn't evasion—it was honesty about a problem that's expanding faster than anyone can measure it.

The numbers explain his caution. The UK has 140 proposed data centres in the pipeline, requiring about 50 gigawatts of electricity. That's 5GW more than the entire country's current peak demand. Friends of the Earth warned that if all these facilities come online, they could consume more electricity each day than the nation does now. This creates an obvious tension: the UK is legally bound to reach net zero by 2050 and aims to meet 95% of electricity demand with clean power by 2030. Adding the equivalent of another country's worth of power consumption doesn't make that easier.

The Hidden Water Problem

While energy consumption dominates headlines, water use tells an equally troubling story that most people never see. A large data centre consumes up to 5 million gallons of water per day—1.8 billion gallons annually. That's equivalent to a town of 10,000 to 50,000 people. Across the United States, 5,426 data centres collectively consumed 449 million gallons daily as of 2021, totaling 163.7 billion gallons per year.

Most of this water evaporates into the atmosphere through cooling towers, never returning to local water systems. Unlike energy, which can theoretically be replaced with renewable sources, water consumption directly depletes local supplies. A 2024 U.S. Data Center Energy Usage Report projects that annual onsite water use could increase by two to four times between 2023 and 2028, reaching 150-280 billion liters.

The water doesn't just disappear—it becomes unavailable to communities, agriculture, and ecosystems. Only 0.5% of Earth's water is accessible freshwater safe for human consumption, making every gallon that evaporates from a cooling tower a gallon someone else can't use.

AI's Thirst

Artificial intelligence has intensified water consumption in ways that weren't anticipated even a few years ago. Scientists at the University of California, Riverside calculated that each 100-word AI prompt uses roughly one bottle of water (519 milliliters). Thirty minutes of AI usage requires slightly more than 0.16 gallons—about one standard water bottle. These figures sound small until you consider scale.

Training GPT-3 required 5.4 million liters of total water. Training GPT-4 consumed 6% of a local Iowa water district's monthly supply. Microsoft reported a 34% increase in global water consumption during development of its early AI tools. By 2027, UC Riverside estimates that global AI demand could drive annual water withdrawals of 1.1 to 1.7 trillion gallons—more than four to six times Denmark's total yearly water withdrawals.

The water footprint extends beyond the data centres themselves. Direct water use for cooling accounts for only about 25% of the total. The remaining 75% comes from power plants generating electricity and semiconductor manufacturing. A single chip fabrication facility can use around 10 million gallons of ultrapure water per day—comparable to 33,000 U.S. households. Producing one gallon of ultrapure water for microchip manufacturing requires approximately 1.5 gallons of tap water.

The Measurement Gap

The industry has developed metrics like Water Usage Effectiveness (WUE), which measures efficiency in liters per kilowatt-hour. The average WUE across data centres is 1.9 liters per kWh, with zero being ideal (achievable only with air-cooled facilities). But a 2016 report found that fewer than one-third of data centre operators even track water consumption.

This measurement gap matters because you can't manage what you don't measure. When Miliband told MPs the climate impact was "inherently uncertain," he was acknowledging that the UK—like most countries—lacks comprehensive data on what these facilities actually consume. The Environmental Audit Committee has launched an inquiry, but they're investigating a moving target that grows faster than their ability to count it.

Technologies That Could Change the Equation

Some solutions exist. Direct-to-chip liquid cooling and immersion cooling can reduce both water and energy usage compared to traditional air cooling with evaporative towers. Closed-loop systems recirculate water rather than evaporating it, eliminating the need for continuous replenishment. Microsoft announced in 2024 its goal to operate zero-water evaporation data centres, though implementation timelines remain vague.

The political response has split predictably. Conservative shadow energy secretary Claire Coutinho argued that barring data centres "will do nothing to tackle climate change, but it will make us all poorer." Donald Campbell of advocacy group Foxglove said Miliband's letter "raises more questions than it answers."

Both miss the point. The question isn't whether to allow data centres—they're already being built—but whether we're willing to account for their true costs. Every email stored indefinitely, every AI query, every cloud backup represents a claim on finite resources. The infrastructure behind our digital lives isn't virtual. It's concrete, steel, electricity, and water—lots of water.

The uncertainty Miliband acknowledged isn't a reason to delay action. It's a reason to demand better data, enforce stricter efficiency standards, and recognize that the cloud has weight. Until we measure the full cost of our digital habits, we're building an infrastructure whose appetite we don't understand and can't control.

Distribution Protocols