Just read @pauldaoust 's “Satoshi Nakamoto and the Fate of our Planet” on the holochain blog. Wow, great article, thanks!
Just a detail, but the graphs got me thinking. I think the energy overhead graph for PoW should be different:
The energy overhead that comes with scaling the network should be, from all I know, roughly the same as PoS. However, the massive difference from doing the PoW does not depend at all on the size of the network of users. Instead it depends on the computing power spent on mining, whereas the lucrativeness of this spending is determined by (mining rewards + transaction fees) * BTC Price. So even if there was only say two users, trading back and forth BTC, the energy spent on mining could be the same or even higher (depending on the prices).
Thus, the graph should start for zero network size to be in the clouds already. Whereas the height of the clouds depends on price, rewards and fees. And then only increase insignificantly like PoS.
What do you think, am I missing something?
However, does not change anything on the points you are making 