This first-google-result analysis[0] from a couple years ago puts the global banking system at around 100 TWh a year which is pretty close to the currently estimated "121.36TWh" for bitcoin[1].
Incidentally, that 2nd article mentions that this is equivalent to the power used by inactive-but-on electronic devices in US households alone.
I guess PC gaming was at 75TWh 5 years ago [2], so I'm not going to worry too much yet.
Total power consumption is the wrong metric; we should talk about the average consumption per transaction. If the totals are close, it means Bitcoin is an extremely inefficient system, since it only processes a fraction of the number of transaction per year that the mainstream financial system processes.
I agree, and I don't think bitcoin will ever be efficient. But I don't think any of us can stop it at this point and it still seems less frivolous than many of the things we spend energy on.
Since energy consumption has historically always increased and since it drives all of human progress and innovation I'm not convinced that the demand side is where we should focus on solving our problems.
Incidentally, that 2nd article mentions that this is equivalent to the power used by inactive-but-on electronic devices in US households alone.
I guess PC gaming was at 75TWh 5 years ago [2], so I'm not going to worry too much yet.
[0] http://climatestate.com/2018/01/15/energy-consumption-bitcoi...
[1] https://www.bbc.com/news/technology-56012952
[2] https://www.forbes.com/sites/michaelkanellos/2015/09/07/the-...