AI Water Footprint Calculator
Every AI chatbot query runs on servers that need electricity and, often, water for cooling. This calculator gives a rough estimate of the water footprint behind your everyday AI use.
What This Calculator Estimates
Running an AI language model takes electricity, and much of that electricity ends up as heat that data centers remove using water-based cooling systems (either directly, or indirectly through the power plants generating the electricity). Independent research, including a widely-cited 2023 University of California, Riverside study on large language models, has estimated that a single medium-length AI query can be associated with roughly a fraction of a milliliter to a few milliliters of water, depending on the model size, the data center’s cooling technology, and local climate.
How the Math Works
This tool multiplies your estimated number of daily queries by a per-query water-use figure you select, then scales that up to monthly and yearly totals. Daily water (L) = queries per day × mL per query ÷ 1000. The three preset options span the range seen across published estimates, since actual figures vary a great deal by AI provider, model, hardware, and cooling method.
These numbers are illustrative estimates for general educational purposes, not a precise measurement of any specific AI provider’s infrastructure. Water use also varies with the season, the region’s power grid mix, and whether a data center uses evaporative cooling, closed-loop cooling, or air cooling.