So the VIM2 (I will soon have a cheaper VIM also) looks good.
Reading throught all the forum, it seems to me that development focus is a bit scattered. So my two cents are this:
Why focus on Android? Yes I understand the need to supply a firmware with Android. However for the price of a hamburger, I can buy one of 100’s android arm tv box. So why wasting your time on this when 2:
Your device is an excellent development and engineering excersise and tool for Linux/UEFI/ miniature formfactor computing etc.
I would focus heavily on opening up the dev work on making Kernel work. Make the detail readily available for one OS (looks like you have chosen Ubuntu so that is fine). I bet the community would run with it. If all the information was available to configure the dts/dtb, and necessary kernel modules (and drivers, but I understand amlogic are probably hard to deal with or licensing difficulties to post their drivers) then I bet before you knew it, you would have people releasing Debian/Redhat/Gentoo/etc etc
I guess that is what Ycto and armbian sort of try to achieve.
You probably already have a business model that means Android-TV. I just think its a waste of time competing with a market that is flooded.
VIM should be more like ODROID. VIM is more powerful that Pi, and to beat them you only need to make it infinitely easy to roll your own linux. (and you already have GPIO)
Android - TV would just be a bonus.