Thanks for your input, and I agree with your observations. The costs and time are major factors inhibiting growth.
However, consider this.
Costs
The current costs for producing parts are high, as you point out. That's because the machines are few and far between. The owners of the machines at present have a small window of opportunity to maximise returns on their investment. The colour printing business was like this, before prices came down and colour laser printers became available to everyone. However, like all commercial ventures, availability and demand will adjust until the price is right. I would bet that in the future your local Kinkos will have a 3D printer. Heck, I'll go one further - your local model club will buy one to generate funds!
Effort
Admittedly producing digital models is time-consuming and hard. However, computer processing power increases all the time, and the CAD software is continually improving to make it easier for even hobbyists to learn the basics. I would suggest that the open-source software development industry is an example of how this could play out. Sure, the professional developers work for the big companies, but there are tens of thousands of developers out there who happily write code that someone might need, for a small donation (shareware). I think there are enthusiasts who would develop digitals cars, and others who would build those cars.
I visualise picking a model to build like you search for a software utility - you go to the sites you know and trust, and choose what you want from a list. Who knows, maybe you even get the source code and do some modifications yourself before sending it off to production!
As for the complexities and costs of moulding and casting, who knows whether it will be necessary, if 3D printers develop in terms of capabilities? In the beginning of their development inkjet printers were laughably bad - now everyone has one and they've become very good (and cheap!).
As I said in the first post - there are millions of reasons why this can't work. The main reason, however, is because we apply the limitations of what we currently know. The thing is, it's what we don't know yet that will make this possible. I didn't know that there is a service to make 3D parts, until now. The fact that there is tells me that the business model for rapid prototyping is beginning to evolve. Now it becomes interesting!