Dish has built a system that satisfies 95% of their customer base, evolving what existed previously (e.g., 3 tuners on a single coax instead of 2), using coax as a distribution system, etc. The choice of what's in a Hopper box targets the majority of customers with a second Hopper picking up another 3-4%. Those asking for 8, 10, 12 or more tuners are a tiny minority and architecting a system to satisfy their needs won't be cost effective for the other 99%. The cost for such systems needs to rest on the 1-2% of the customers that need it, so saying a Hopper should have 6 tuners is unrealistic. Where Dish dropped the ball is in not integrating two (or more) Hoppers so that the users don't see a hard line between them: where material is located, which receiver has a spare tuner at the moment, ... This could be fixed but Dish has chosen to focus software development on things like AutoHop, Netflix and a new skin on their U/I. They seem to want to attract new customers with glitz rather than keeping existing customers with a system that works as promised at an attractive price.
I'm thinking the traditional approach to send all transponders on a given satellite to the receiver where one channel from one transponder is extracted may not be the best approach to a new system. Why not move the tuners closer to the receiver then send the much lower data-rate to the places where the channel is viewed or recorded. If MOCA limits what can be done in a system, choose something else (like 10G ethernet). Does the place where material gets recorded physically need to reside next to one TV or is there a better location that makes expansion and distribution to viewing locations easier? Does a different partitioning of the system reduce or increase cost for the single-TV customer? I don't have the answers, just tossing this out for what it's worth.