Judging how successful somethings is based on how popular their messaging service is seems a bit counter-intuitive. During my life I have seen IRC, then ICQ, then AIM, then MSN, then Facebook hold the top spot for IM client over period as low as 15 years. What more, at this moment all of my friends are reachable on gtalk simply due to the rising popularity of the Android platform, so who's to say how things will change going forward.
When it comes to communication people simply use what's most convenient, and has the bigger percentage of their friends. Obviously this will shift and change as services gain and lose popularity, and I think the past couple of decades have illustrated beyond a shadow of a doubt that eventually everything loses popularity on the internet (save perhaps trolling). Facebook is already a gigantic corporation with thousands of employees, millions of line of code written to support their model, and an ever growing turnaround time for new features. Obviously that can be very useful for company with such a large size and scope, but it is not without problems.
All it would take to drown Facebook would be for a new service to come along with some sort of game-changing idea; at that point the monolithic structure of Facebook will prevent them from adapting quickly enough and they will begin to lose their early adopter crowd, and with them more and more of their newcomers. Sure, they will keep much of the mainstream crowd they managed to attract over the years, much like myspace still holds a huge share of their original market segment, but that will be inertia more than anything.
This is not some wild speculation either. We've seen this happen several times already in the short history of the internet, and we will see it happen again, and again, and again for as long as the internet exists. This is just the life cycle of a web based company. So in other words Facebook isn't really winning any war, because honestly, there really isn't any war. At best it is yet another period of reshuffling during which yet another service ends up on top.