The way its looking right now in the voice/video messaging scene is we're gonna have communication islands and the only way to reach someone is to run five different service's clients.
Is A/V over XMPP standardized?
Even so, I suspect it will be a long time before Skype, Apple, and everyone else does anything but try to set their own standards.
There are pragmatic reasons for the fragmentation we're seeing, although those don't necessarily correspond to the customer's favored outcome. Federation reduces the stickiness coefficient of your product, and forces you to design your application around a protocol that may not fit your needs ideally, cf. Skype's supernode architecture.