Filed under sustainability.
Every data center spends a meaningful share of its power moving heat. How large that share is gets decided earlier than most operators assume: not by the cooling vendor, but by the climate the building stands in. A site in the Nordic north spends most of the year at temperatures where outside air does the work largely by itself.
The second climate decision is water. Evaporative cooling trades water for electricity, and in a catchment that is already stressed that trade eventually meets a regulator. Siting where the climate allows dry cooling for most of the year removes the question entirely, which is why water strategy appears in our screening model next to grid capacity.
None of this can be retrofitted. An operator can upgrade chillers, but nobody can upgrade the outside air. Which is the quiet argument under all of our siting work: the most consequential engineering decisions on a data center are made while it is still a forest.
