It certainly is a valid point. However, your assumption seems to be that the biggest barrier to entry would be hardware? If so, I disagree, with all due respect. I spent best parts of the summer learning iOS - it's not extremely hard, but it has many non-obvious corner cases that you'll need to spend some time to discover. So with almost any reasonable hourly rate (even at the minimal wage) I've invested much, much more in the learning of the framework that I would spend on a minimal hardware that will allow me to develop for iOS.
Let's face it - if you're serious about porting your apps to iOS, you would buy a mac at some point.
It is hardware and software, it is the Mac ecosystem; at least for me. My core business is the Web and we are getting into Android. I already spent several thousand dollars in Hardware. If I want to develop in Mac I need a Mac. It is not as easy as Linux. I could of course built my own Mac, but still it gets pretty expensive. Specially since iOS is not our core market.
Now, if iOS ends up being big for us, then yeah, we would stop using the service and buy our own machines. But then the service served its purpose.
I guess what I am trying to say is that this kind of service has a market and is not that small. But I do agree with you 100%. If you are serious about iOS development, you need to get your Mac.
Let's face it - if you're serious about porting your apps to iOS, you would buy a mac at some point.