They need some sort of breakout box for the convenience alone. You'd be re-connecting the HDMI everytime if they hadn't provided it, even if they provided some kind of simple repeater in some SKU's for those that don't want the TV views, it would be again an added type of SKU which will probably be even more complicated (Those with camera, those without, those with move contollers(?), those without, and now with the breakout box, without the breakout box).
Making it optional would also mean further splitting a what is going to be small user base at least definitely in the first years of the products life, and would be detrimental for devs who want to make use of the "social" aspect of the product, and sharing feed of the game that's viewable by everyone in the room is one of product's defining features so far.
So I don't even consider this a possibility, to not have that box.
Edit: Dammit guys, it's not a secret sauce box, forgot what this thread was. The box needs to have some processing power as it needs to format the fish eye view into the TV view, and do this flexibly (different settings for different games), and it has to relay a second view if the dev chooses to do so (entirely different view from the 3d view sent to the headset), and do this without adding a hell lot of latency. So it's not a dumb box, but not secret sauce GPU.