Oh please. You have no clue what you are talking about. I am a hardcore UNIX admin with about 15 years experience. Name a version of UNIX and I have done it. Solaris, SunOS, IRIX, OSF1, Ultrix, NextStep, HPUX, AIX, FreeBSD, OpenBSD, Linux, etc.. If it has been used in the past 10 years, I have at least hands-on experience with it.
The decision to use an O/S has nothing to do with the cost of the O/S itself. The actual cost of the software is meaningless. Yes, if you are putting it on every desktop it becomes an issue, but when it is used on a server, it is a very small piece of the pie.
The major costs come from hardware and the people needed to run it. When picking a solution, if the company is smart, they will go with a solution that is not going to become some legacy O/S in a few years. They also need to take into account how difficult it will be to hire/train and retain people to run the systems. One of the main reasons Linux is so popular, aside from the fact that it works, is there are a lot of people out there who are familiar with it and there is almost no chance the O/S is going to go into the dustbin anytime soon.
I absolutely love Solaris, but can you guarantee it is going to be around in 5 years time? If you asked me that 5 years ago, I would be 100% sure it would be. Now, I don't know. My gut feeling is it will not be. (SUNW is currently trading at $4.17 a share.) Then what happens to everyone who is using it and has to transition to another platform?
I have worked in shops that have transitioned from Solaris to Linux. The cost of the O/S had absolutely nothing to do with it.
But since that time, IBM decided to (likely illegally, case is in US Federal court now) port Unix features to Linux and give them away for free, and since that time we've seen an explosion of Linux worldwide since the foreign governments could now obtain this software completely without cost, at the expense of the US Unix vendors. Just because you weren't aware, doesn't mean it wasn't happening.