Lyn Margulis thinks they don't though, and she's just a hair more qualified than me to judge. She says they change genes too readily to be consider genetically isolated, and should be ID'd as strains instead.
All I can say is that the species concept is a little fuzzy too with eukaryotes (ring species, anyone?), and bacteria maintain a meaningful separation between their "strains" so the concept is close enough. The reason I want them to be separate species is because I still want to demonstrate speciation in a petri dish. Maybe someday.
(And I do recognize that the fact that I want bacteria to have separate species has no influence on whether they have separate species. Unfortunately.)