Well isn't there some practical limitation to the MFH-2 system approach which is why MFH-3 was designed in the first place?
Or is their something maybe new about "D2" over the former SWiM multiswtich based MFH-2 system that allows for covering the very large apartment complexes?
Maybe MFH3 was designed because it was cheaper for very large installs than MFH2, or at least was planned to be?
I don't know a lot about MFH3, but from the sound of it there was a very expensive head end that essentially tuned very many (or maybe all?) channels simultaneously, so the MPEG streams could be sent out over ethernet to the receivers. A high end DSP could trivially 'tune' an entire transponder's worth of channels, so you just need a box full of them to tune everything (but I have no idea if that's how they implemented it) The receivers had nowhere to plug in coax, only ethernet, and had no tuner. If you made them in similar quantities, they'd be less expensive to make than a regular receiver because there would be fewer parts. It sounds like the H20i was similar in many ways to a Genie client using RVU.
Since any modern building is going to have ethernet running everywhere anyway, if your building is large enough it is probably cheaper to have one really expensive head end and that's it. You're using the ethernet infrastructure you already have in the building. There's no coax running anywhere, no amps, no switches, no worries about tuner limits in the units. If they did it today you wouldn't even need receivers if it could be made compatible with RVU, though you'd need RVU clients capable of recording unless you connected that head end with a deduplicating disk array so it could act as a DVR for the entire building. Gigabit ethernet could carry 500-1000 HD channels, and 10Gb ethernet is very affordable as a backbone these days. If they used multicast, which I assume they would, you'd only need to carry each channel once no matter how many people were watching it.
It really makes me wonder why they abandoned MFH3, but if I had to guess either they couldn't make it as cheap as they planned, or maybe the point where that become cheaper limited its market so much that it couldn't support the cost of continued development. Perhaps if they'd held out long enough to make it compatible with RVU so they didn't need specialized receivers MFH3 could have become more competitive over time as technology drove down the price of that head end.