Playing devil's advocate here: to me, this doesn't prove Bitcoin doesn't consume a lot of energy. If anything, the same energy consumption as about 200 million refrigerators is a lot (and the average American refrigerator is a lot bigger than they are in my country).
[~snip~]
The numbers are no doubt true, but it doesn't mean anything. You can't use transmission losses to power a Bitcoin miner, and you can't avoid the losses either. It's a trade off between power line thickness and power losses.
This is not meant to tell Bitcoin doesn't consume. It does. This was meant to compare with things closer to the American "average Joe", so it doesn't look as scarily huge like <insert country name here>'s total consumption (where industry consumption may or may not have a major role).
Of course, you are not wrong, but there is the volume comparison and there is the utility comparison. It looked interesting to me as volume (while you've discussed the utility). However, if this is not useful, no biggie.
PS. The consumption per transaction is a bullshit metric and maybe that should be debunked/explained too (but not to me and in terms a non-bitcoiner can understand).