The article you linked doesn't do much to refute "data centres are acoustic weapons", just insists that symptoms are explainable by audible noise pollution. Sure, that makes it better?
A) don't exist, therefore not a concern
B) known to exist, therefore not a concern