I've written and launched two mobile apps on the iphone over the last year and both have had success in economic terms. One app is free and tied to a web service and it has a significant impact on the popularity and number of users for the web service. The second app is a paid app - and I can tell you that it is producing some actual revenue, enough that if I was a solo developer it'd be paying my bills.
That said I think that if you're launching a company for mobile products you don't want to put all your eggs in one basket. So either support multiple platforms or aim to have multiple products on your main platform.
I think there is big potential in Android but at the moment it is totally unproven as a platform where you can actually make money (please point out some info on this if you have any I am really curious about the economic potential of Android).
Blackberry is also interesting since pretty much everyone I know who's under 25 has one, but it is a platform where selling apps doesn't seem to have caught on that well. I've discussed it with some heavy blackberry users and apps are not something they really care that much about. So you'd want to try to find out some numbers regarding Blackberry app sales.
In the end it depends on your target market/product.
Are you building an enterprise targeted mobile app? - Build for Blackberry first and perhaps iPhone next.
Do you want to launch one consumer focused mobile app with a large feature set and perhaps some web service integration? - target a few platforms and make it available to as many users as possible.
Are you trying to build a series of small purpose built apps? - Definitely start with iPhone and get some revenue first.
My 2 cents.