The litmus test for a tool
A good tool makes it easy to do simple stuff and possible to do complex stuff. TFS, on the other hand, makes it simple to do extremely easy stuff (the kind you see in a 10-minutes tech demo), moderately difficult to do simple stuff, and next to impossible to do complex stuff. One example that pretty much made it for me: it's surprisingly painful to work with files in TFS unless those files are part of a Visual Studio project. So if you use *any* file that's not a good fit for VS (say, a folder containing plain text documents, or a small Python script), prepare to suffer.
As a versioning tool, TFS is painfully behind the curve compared to most (free) alternatives. It will basically cripple your productivity. As an Agile project management tool, it all but guarantees that you'll either end up ignoring it, or you'll never do Agile properly.
In conclusion, the only decent use case for TFS is a Company that doesn't yet use any versioning system, and has a religious dedication to only use tools produced by Microsoft. It might also be an improvement for people using ClearCase. For all other cases I can think of, TFS would be a net negative even if it were free. The fact that it costs serious money adds insult to injury.