1 Probably economy of scale. I bet the DAS is a special order, which has to be sheperded all the way through the production line, whereas the server is off-the-shelf.

2 I don't know which DAS you were looking at, but if it's the one we build which they rebadge (or heck, most any DAS)... there is quite a bit more to the thing than just a stack of drives.

Dual redundant smart power supply modules, ditto fans, redundant enclosure services modules, redundant internal loops, lots and lots of R&D.  It is far more likely to choke itself on some silly internal coding error, than it is to lose your data due to hardware malfunctions... and it is very very unlikely to trip itself on coding errors.  Not that they're not there... just that they exist primarily in remote boundary conditions.

Also, the drives will have special firmware, and possibly be the pick of the litter, so to speak.

3 I don't think Dell do special drive versions, though enterprise-class drives certainly do have different firmware than commodity drives.

It does offer redundant controllers as an option, though I didn't configure that.  Just having the option increases the complexity and price, of course.

The most expensive component, though, is the drives - Dell's drive prices on their US site are hugely inflated.  They're actually about 40% cheaper if you buy them in Australia (though in Australia the rest of the server is more expensive).