Those numbers I gave were from your benchmark where Max Q draws an average of
6 watts, the CPU draws an average of
0.45 watts and the GPU is drawing an average of
2.1 watts. The Max power draw in the test for Max Q is 8.1 watts for the system, but this is peak consumption. Switch docked with lower clocks isn't close to any of this, with a power draw of 11 watts, yeah there is certainly more going on with the Switch, but the difference here is closing in on two times, and that is substantial enough to say there is a difference in 16nm vs 20nm... I mean we know that the SoC consumes 4 watts less with portable clocks and the screen being on, and that 7.1 watt power draw I'm talking about is only 1 watt less than the Tegra X2 system's peak power draw.
I don't know how to have this conversation with you, but I don't think your idea is really valid, not enough to shut down
ilikefeet's speculation about a passively cooled Switch on 12nm with portable clocks.