The role of ProQuest for dissertations is not a "publisher". Even if you "publish" your dissertation with them, it is still considered "unpublished" work in many fields. And you can go on to publish exactly the same content in other forms, e.g. as a book or divided into papers. In this sense, they are like a preprint server.
For dissertations, ProQuest is a repository. The main service they provide is giving access to dissertations that would otherwise be very difficult to locate. Traditionally, they did this by keeping huge archives of microfilm; now I suspect it is more digital.
In my experience, there are only two long-term archive systems commonly used for unpublished theses in the U.S.:
1) The institution itself usually keeps a copy in the library, perhaps on microfilm. More recently, institutions have begun moving to an electronic model, often called "ETD".
2) ProQuest, which has gone through many name changes but was founded in 1938.
Personal webpages are great, but they are not really a long-term solution: who can guarantee an academic's work will be available on a personal web page after they retire? A few fields have sites such as arxiv.org which have a potential to be long-term repositories. But otherwise the options above are about it.
So, why send your thesis to ProQuest? It's a personal choice. But doing so does help ensure that the thesis will be available if someone wants to read it in 50 years' time, when you plan to be retired on a tropical island without email.