The only viable free software business models I've witnessed are of large IT companies (e.e. Sun, IBM, Nokia etc), where they can do quality middleware and even end-user solutions as free software. This is quite limiting, since naturally there isn't a lot of those companies.
A lot of smaller free software/open source projects, on the other hand, are quite crappy -- since understandably they're done as volunteer work without proper resources to make something polished and finished. So we're generally torn between two options: 1) use the perpetually unfinished software, or 2) suffer from the intentional misfeatures of the proprietary software. Therefore I tend to think the demise of proprietary offerings isn't coming soon. Free software does seem to improve over time, though. Unfortunately, the process takes a lot of it.