I think the microconsole will be more powerful than the handheld variant but it won't be a ridiculous gap in power. That would undermine the whole point of the scalable/sharable games.
I suspect they will have one processor core (system on chip) which is an ARM CPU + AMD mobile GPU part. This is the heart of both systems. One of these will be able to pump out Wii U-ish graphics at about 4 watts power output (thus suitable for a tablet style handheld).
The microconsole IMO will use pretty much the exact same processor, it'll just have 2 or 3 of the SoC cores ... so 2x-3x the Wii U at a very low 8-12 watts output. Handheld has 2GB RAM LPDDR4 RAM, microconsole has 4GB LPDDR4 RAM (same stuff Apple uses in the iPhone/iPad). No disc drive, no fan, no giant power brick, scales down in cost extremely quickly. Both handheld and home variant can share the same physical carts or you can choose to buy your games digitally.
I don't think Nintendo dares release a console again at above $200 either. Unless they have a magic Wiimote like gimmick, they just have a terrible time selling anything at a premium price point. Profitable at $200 from day 1 is likely a big priority. That way even if it sells only 20 million, they'll still make money on it.
Competing with the PS4 is a non-starter. PS4/X1 will be around 80-90 million install base by 2017 (even worse for 2018). No one is going to buy some "Nintendo PS4" showing up that late when Sony/MS have established online communities, hundreds of games, mindshare, etc. etc. It'll be the Wii U all over again where no one cared that Nintendo finally caught up to PS3/360 graphics, like big deal. Too little, way too late.
Something cheap, that plays their entire library of games, can interface with amiibo toys, profitable from day 1, I think will be Nintendo's play (not saying I personally agree with this approach, but I am just looking at it from their management's POV).
