Water resource management is complicated enough with things like industrial runoff, farming runoff, and all of these kinds of issues without then placing data centers into the mix as well. And this is an area that is going to become an increasing problem as AI use and generative AI use grows.
Water anywhere is not the same as potable, drinkable water that comes from, say, a reservoir or from an aquifer. And as we said in the introduction, these are finite sources. So while 71% of Earth is water-covered, only about 1% of all water is actually drinkable. The majority of freshwater is actually just frozen in the ice caps. So we are looking at a very valuable resource.
I know that the OECD AI Policy Observatory has said that in 2027 global ai demand could drive as much as 4.2 to 6.6 billion cubic meters of water withdrawal. Just to put that in context, that is half of the UK's annual water withdrawal. So that's just an incredible amount of water just for global ai demand.
Collection
[
|
...
]