3d Printing for low production run?

Can it be cast?  If it can be cast, get it done in lost wax.

The part is too complicated for casting, I would have to make it from several pieces and then glue it together. The problem is that there are internal cable channels etc. I'm not saying it wouldn't be possible, but just not very practical. I use the solidoodle for prototyping and it works great, but I also have to glue the parts together. I was looking for a printer that is powder based or so, which would allow me to print the part as one piece. I'm just not sure how strong those parts are and if there are any other potential issues if used for finished parts.

I think renting a printer might end up to be cheaper. I'm also thinking about a few other potential products I might use it for.

I'm a little worried about posting anything at this stage. I know it is unlikely that it gets stolen, but there are a lot of mostly Chinese copy cats in the market that I am in and I just don't want to risk anything.
To just give you a better idea:
It's 5 separate parts each of them varies in terms of complexity. A medium complex part for example is a battery housing that will clip onto the main part (just like a canon camera batter clips onto a charger). The housing contains the actual battery and then has internal cable channels that lead to the contacts. Same goes for the main part. There are three cable channels that turn 90 degrees three times. The main part also has a shape that can't be molded very easily. It is basically a U shape that has an L shape attached to it at a 90 degree angle (U is in the x/y plane and L is in the Z plane). Cables run from both ends of the U internally all the way to the end of the short end of the L. Sorry for this complicated explanation, I hope it makes sense.

I just saw it happening a few months ago on another message board. A small european company posted some preview pictures of a new product that they were going to release within about 10 weeks. Within less than a month, a Chinese company released an exact copy of the same design. It was poorly machined, but for a 4th of the price the original designer was going to sell them for. It really made me think twice before posting something. Especially if a high percentage of the selling price goes into R&D.

I definitely won't print it on a consumer printer. I was thinking more of something along the lines of this: http://www.stratasys.com/3d-printers/pr … rtus-250mc
I just don't know how practical it would be and how durable the parts come out.

Of course I will offer pics etc of it, but not until it is ready for sale. I already have re-sellers on the east cost and west cost for a couple of my products as well as my own store, so there's no point in giving away most of my profits to a company like Shapeways (wouldn't work anyways, because there's a lot of assembly needed, like electronics etc.). I'm not saying it is a bad service or anything, it just doesn't make sense in my case.

I just saw a high resolution picture that was printed on one of their FDM production machines. Looks pretty disappointing, not much better than what I get in my sub $1000 solidoodle, or am I missing something?

Ian, do you know how many mm/sec these fortus machines print at?

I noticed that the black region on those industrial printed pieces has a pretty terrible top surface finish.  It's funny to see those problems on an industrial machine when we have the same issues on some of our Solidoodles.  The difference is that we have the flexibility to easily tinker and address those problems directly.