The analysis of Alex de Vries contains huge methodological problems.
#1:
And while he does have confidence in his estimates, the problem with this method is that these manufacturers
are extremely secretive. "Sometimes the best information we've got is really shaky eyewitness accounts. That's the stuff we have to work with," he says.
Does this look like reliable data to work with? I wouldn´t take estimations that
are based on "shaky eyewitness accounts" seriously at all. These eyewitnesses might
be completely biased, which would obviously distort the results of the whole
study (e.g. if you talk to a disgruntled ex-Bitmain employee he might not be 100 % honest,
because he might still have a grudge against his former employer).
#2:
For the majority of the network no information is
available at all. At this time, it therefore cannot be ruled out that hashrate
simply does not reflect a large part of the electricity consumed in Bitcoin
mining.
While this seems theoretically possible, it is strongly likely that the hashrate is indeed
a pretty good indicator of the electricity consumed in Bitcoin mining.
There are several other flaws in his method (e.g. he bases some of his calculations on the
lifetime energy requirement of an Antminer S9 while completely ignoring the possibility
that more efficient ASICs emerge).
Besides, he completely ignores the fact that some energy usage might come from sources
like the geothermal renewables in Iceland or that the energy providers are actually happy
that miners are making use of excess energy in the system.
If anyone wants access to the full article, send me a PM (not sure if I´m allowed
to link it here, because it was published behind a paywall of a journal).