Amazon revealed that its data centers consumed 2.5 billion gallons of water globally last year, equivalent to approximately 5% of metro Seattle's annual consumption. According to reports from Bloomberg, the world's largest cloud-computing company disclosed this figure as part of its sustainability report, showing that it cools server farms more efficiently than some major tech industry peers. The company reported a water usage effectiveness of 0.12 liters per kilowatt hour, down from 0.15 liters in 2024, and below Microsoft's 0.27 liters per kilowatt hour. As reported by The Verge, Amazon also disclosed that it returned two-thirds of that water back to communities it operates in, mainly through investments in public infrastructure projects as part of its broader goal to be "water positive" by 2030.
Amazon's water efficiency metrics position it favorably against industry competitors. As reported by CNBC TV18, the company's 0.12 liters per kilowatt hour significantly outperforms the estimated industry average of 0.84 liters per kilowatt hour. Microsoft's water usage effectiveness was 0.27 liters per kilowatt hour in its most recently reported fiscal year. Amazon's efficiency figure is based on sites that AWS owns or leases, excluding colocation facilities that accounted for about 20% of Amazon's computing power in 2024. According to The Verge, Amazon reported Thursday that its data centers on average used 0.03 gallons of water per kilowatt-hour of electricity in 2025, representing a 52% improvement in water efficiency since 2021.
According to CNBC TV18, AWS data centers use air-cooled systems that chill servers with air blown in from the surrounding environment. On days when temperatures exceed 85 degrees, this air is pushed through water-soaked filters, cooling it on the way into data center halls and evaporating some of the water. Facilities in water-scarce regions like the Phoenix area and Saudi Arabia use air-cooled chillers that circulate refrigerant and water in closed loops without relying on outside water sources. Amazon's water usage effectiveness calculation includes both evaporated water and liquid discharged to wastewater systems. As reported by The Verge, Amazon has tested and confirmed that it can operate buildings at 85 degrees Fahrenheit, reserving evaporative cooling for the hottest days of the year. The company evaporates water less than 10% of the year, and has found ways to use less energy at data centers, which indirectly reduces the amount of water needed to cool power equipment.
A new study by Cambridge-led researchers reveals that AI data centers are creating significant environmental impacts beyond water consumption. The research found that land surface temperatures around AI data centers rise by an average of 2 degrees Celsius (3.6 degrees Fahrenheit), with some areas recording increases as high as 9C (16.2F). Researchers have called this the "data heat island effect," with temperature increases being felt up to 10km (6 miles) away from facilities. The study, using NASA satellite data, measured land surface temperature globally from 2004 to 2024 and cross-referenced it with more than 11,000 AI data center locations worldwide. More than 340 million people living within 10km of a data center could be affected by these temperature increases, which researchers describe as having a "remarkable influence on communities and regional welfare."
The global landscape for AI data center construction is experiencing unprecedented acceleration, with more than 11,600 data centers active worldwide as of June 2026. Most data centers are located in the United States, which has more than 4,300 facilities, according to Data Center Map. Europe is the second-largest hub, led by the United Kingdom with more than 540 facilities, followed by Germany (520+) and France (390+). Across Asia, China (360+) and India (300+) lead the region, while Southeast Asia is expanding rapidly. According to Synergy Research Group, the number of hyperscale data centers worldwide has nearly doubled since 2021, from 700 to 1,297. Global investment bank Goldman Sachs expects a combined $5.3 trillion of capital expenditure between 2025 to 2030 for the four largest hyperscalers: Microsoft, Amazon, Alphabet and Meta. Major upcoming projects include Meta's $27bn Hyperion campus in Louisiana, Microsoft's multiphase $20bn data center campus expansion in Wisconsin, Amazon's $25bn investment in Mississippi, and Google's Project Spade in Missouri.
