Let the numbers settle down for a moment and observe what's happening in the GNU/Linux world. As a matter of fact these numbers are not random, they all depend on real people making real decisions and undertaking concrete actions. A review of these initiatives should give us a better idea than the statistics alone.
Starting with Ubuntu, the most popular distribution to date, the near future seems bright. Mark Shuttleworth said that Ubuntu aims 200 million users by 2015 (at minute 6' of the Ubuntu Keynote may 2011). Is it doable? What does he has in mind? My guess is that he has a plan and he didn't say that just for fun. In fact, we can read things going on at Canonical Ltd. that give some weight to Mark's words. The Spanish region of Andalusia are deploying an Ubuntu based platform in 220,000 computers of schools throughout the region by 2012, the French National Police will switch their 85,000 computers to Ubuntu by 2014 (ubuntu.com), 10,000 computers of the German group LVM are now running Ubuntu (canonical.com), Ubuntu is available from factory on the ASUS Eee PC (canonical.com), Dell is known for selling some machines with Ubuntu but the matter is not clear. Well, this does not make all the 200 million but it's a great start. More than anything this is an exposure of Ubuntu to millions of people that might chose to use it in their home computers as well.
Red Hat Inc. is doing very well too. It has acquired Gluster to be up-to-date with cloud computing and has a total equity of $1.5 billion (Business Insider, Wikipedia). Among their customers we can cite DreamWorks (entertainment), CEPSA (oil company), European University Institute (education), Mazda Austria (automobile), Brazil's Ministry of Health (government), SFR (telecommunications)... and the list goes on and on (redhat.com).
There is an amazing documentary by BBC that covers some of the uses of GNU/Linux in various situations, especially in developing countries, showing the incredible potential of the platform.
I foresee two major events in the future that could completely change the rules of the game. The first is the much discussed end of desktop computers and shift to portable devices and cloud computing. This would simply be a change of mindset for everybody and every platform. Maybe this time the development of GNU/Linux can start the run together with the other platforms and eventually win. Although the cloud looks like the hot trend of the moment, I'm not enthusiastic about it at all (it means handing your private information to third parties...).
The other event that might happen if GNU/Linux desktops start gaining market shares, is that companies of proprietary OS will start releasing some versions of their OS for free or even under a sort of 'open-restrictive' license. This is a little bit what happened with Windows XP lately. XP still has an incredible market share considering that it was released a decade ago and is not supported anymore. This is mainly due to the Vista flop and Microsoft giving XP for free to those who want to downgrade (zdnet.com). Microsoft already did that sort of thing back in 1996 when they released Internet Explorer for free to “cut off Netscape's air supply” (wikipedia.org). Apple also gave Maverick for free to anyone that wishes to upgrade (and ended support for older OSs). The difference between the browser war and an OS war is that FOSS is not developed by one single company, they are made by people, anyone, and this turns upside down the old model of software development.