Earlier quoted context omitted.
It's software, it can be optimised much quicker than optimising the politics and dynamics of the existing banking system.
Energy consumption goes up with mining difficulty but its throughput stays the same. This manta keepa being repeated by the bitcoin crowd as if it were a fledgling technology, when its already 10 years old. Is mass adoption just around the corner, do you reckon?
There's lots of people working on more energy efficient, yet still secure cryptocurrencies, but IMO we won't see truly energy efficient cryptocurrencies until we shift the trust away from the central chain and into individual actors.
Theoretically you should be able to establish a trust coefficient for an action given the signature of the actor & the gossip on the network about their outstanding debts, their current balance & their non-transferable value.
You'd still need to pay the cost for obtaining the information about the actor and a method for verifying that you haven't been subjected to a sybill attack w.r.t. that information, but that amount of data transfer/computation should be orders of magnitude lower than the current PoW technique that bitcoin uses.
There's no way that the rough system I am talking about would be adopted by bitcoin, and I'm not gonna be working on it, I haven't seen a project that looks similar so I'm guessing it will still be quite some time.
If I was going to go with a cryptocurrency that is energy efficient and currently exists I'd probably take a serious look at https://www.nanode.co/