Amazon's data centers used 2.5B gallons of water in 2025, claims 7x efficiency lead
Amazon finally reveals water usage data – 2.5 billion gallons and 2% less than 2024 despite growth.
Amazon has released its annual water usage data for the first time, revealing that its global data center operations consumed 2.5 billion gallons of water in 2025. The figure represents a 2% decrease from 2024 despite ongoing expansion, with a water usage rate of 0.12 liters per kilowatt-hour. Amazon claims this makes its data centers seven times more water-efficient than the industry average, citing a peer-reviewed research paper for the benchmark. The company attributes its efficiency to a cooling strategy that relies on air cooling about 90% of the time, only switching to evaporative water cooling on the hottest hours of the hottest days. Amazon also says it has raised server temperature tolerances to reduce cooling needs.
The disclosure comes just after Seattle enacted a one-year moratorium on data center construction, a move some Amazon employees publicly supported. In its report, Amazon compares its water efficiency to rivals Microsoft, Google, and Meta, showing each using more water per kilowatt-hour in recent years. However, the comparison is imperfect: Google's data appears to focus specifically on Gemini AI data centers, while Amazon reports across all operations. Amazon also excludes indirect water use from power plants supplying its electricity and from data center construction. The data highlights growing scrutiny of AI-driven data center resource consumption, with water and energy use becoming central to regulatory debates.
- Amazon's data centers consumed 2.5 billion gallons of water in 2025, down 2% from 2024 despite expansion.
- Amazon claims a water usage rate of 0.12 L/kWh, seven times more efficient than the industry average.
- The company uses air cooling 90% of the time, reserving evaporative cooling for extreme heat; it does not factor in indirect water use from power plants or construction.
Why It Matters
Raises the bar for transparency and efficiency in AI data center resource debates amid growing regulatory scrutiny.