To at least partly answer my own questions:
The handbrake test show 100 Watt DC load power consumption. I want to build a mini itx system with Trinity and use it for gaming (i.e. high usage of both GPU and CPU unlike in a HTPC).
Do you think a Pico PSU rated at 96 Watt (8A) at 12V would be problem since the power consumption is probably close to the PSU's rated maximum? (Pico PSU 160 + 120 Watt power brick, one SSD and one fan in addition to mainboard, cpu and cpu cooler)
I misread the power consumption number. The 100 Watt figure in the article is with a dedicated graphics card. The Prime95/FurMark figure on page two shows 92 Watts worst case DC power consumption. This should be fine. The 120 Watt Pico PSU version rated at 84 Watt at 12V might be operating outside its specifications in this worst case scenario, though. (I have no idea how much of these 92 Watts is on 12V)
Does anyone have voltage information to compare the 5700 and the 5800K? I have seen AMD processors undervolt very well and I would like to see how AMD achieved the lower power consumption with the 5700.
Tom's hardware reduced the VCore of a 5800K to 1.275V on stock clocks (default might be 1.3625V ) which yielded a reduction of average power usage of ~14 Watts.  ComputerBase shows 20 Watts less for Cinebench with a VCore reduction of 0.15V. 
More interestingly, a user at HardwareLuxx reduced the clocks to 3.4 GHz with a VCore of 1.14V (default: 1.416V) and got a Prime 95 power consumption reduction from 122 to 88 Watts (no GPU load) . This is quite a big change to the VCore for such a small change in clocks!
I would love to have more data since every Watt counts in a small case such as the M350. Still, I'm intrigued now and might just go for it as soon as NH-L9a compatible mini-itx motherboards are available in my part of the world.
 http://www.tomshardware.co.uk/a10-5800k ... 36-11.html
 http://www.computerbase.de/artikel/proz ... u-test/13/
 http://www.hardwareluxx.de/community/f1 ... 19145.html