There's a high correlation between geography and social graph, while the part of the social graph that gets "cut" will give an incentive to make interoperabilit

There's a high correlation between geography and social graph, while the part of the social graph that gets "cut" will give an incentive to make interoperability work, as people will care about it. So your premise is true(ish), but the conclusion can't be drawn from it (and your comparison falls apart) because we don't need strict dependency on geography, only a (relatively strong) correlation.