Amazon Reveals Data Centre Water Use Reaches 2.5 Billion Gallons

Amazon has revealed for the first time that its global data centres used around 2.5 billion gallons of water last year. While the figure is high, the company says its systems are still more efficient than the wider industry.

In a recent blog post, Amazon claims its data centres are “seven times more water-efficient than the industry average”. The company highlights efforts to reduce water used for cooling, especially as AI infrastructure faces increasing scrutiny over environmental impact and pressure on local water supplies.

However, earlier reports paint a more complex picture. A leaked 2022 memo suggested Amazon’s data centres could reach 7.7 billion gallons annually by 2030. Meanwhile, a 2023 report by The Guardian claimed water use had already hit 10.5 billion gallons in 2021. That same report alleged Amazon had discussed limiting public awareness of its water consumption.

According to SourceMaterial, Amazon operated 924 data centres worldwide as of late 2024, with further expansion expected as demand for AI services grows.

Amazon does not directly address these earlier figures in its latest post. Instead, it says the 2.5 billion gallon figure is based on utility meters and verified by third-party auditors. The company also attempts to put the number into context, noting that Americans use around 3.3 trillion gallons of water annually for lawns and gardens – over 1,300 times more than its data centres.

The company says it is taking steps to reduce its reliance on fresh water. These include:

• Using reclaimed water from wastewater treatment plants instead of drinking water
• Running 26 facilities entirely on reclaimed water, with 130 more planned globally
• Relying mainly on air cooling, with water used only during the hottest periods

Amazon has also set a target to become “water positive” by 2030, meaning it aims to return more water to communities than it consumes. It says it has already reached 75% of that goal.

Despite these claims, questions remain about how Amazon compares itself to competitors. Critics point out that its efficiency figures appear to combine traditional and AI data centres, while rivals such as Google report water usage specifically for AI-focused sites. This makes direct comparisons difficult and potentially misleading.