And banks don't have any mining at all, so they don't have to spend any energy on that.
The absence of a need for mining gear presents with the problem of no provable backing for fiat currencies besides faith in the system. The gold standard used couple of years ago served as a means of backing up currencies by real value and was adopted by many countries, but has long been dropped.
Bitcoin's algorithm offers such proof of value while still being environmentally friendly. I would call that an improvement.
And right now Bitcoin only covers how many, 700k transactions per day? This is nothing compared to Visa. Until LN is widely adopted, Bitcoin network is less efficient at per transaction basis.
I doubt the ability of Bitcoin to scale up to certain levels, but it is a growing technology with lots of room for improvements, so I wouldn't want to nip it in the bud. A couple of BIPs aimed at improving scalability has already been implemented and some like Schnorr and Taproot are yet to be.
Also, enviromental impact of both banks and Bitcoin is tiny compared to industries like transportation or agriculture. Even if they were 100% green, it still wouldn't make any difference.
This is an excerpt from the article I linked in the first reply:
"...This could help turn loss making renewables projects
profitable and in time as the industry matures and
settles as permanent in the public eye could act as a
driver of new renewables developments in locations
that were previously uneconomical."