@bca3d3cb According to some estimates, the energy cost of a Bitcoin transaction is about equivalent to that of two billion Visa transactions: https://www.marketwatch.com/story/heres-how-many-visa-transactions-can-be-completed-using-the-energy-to-mine-one-bitcoin-11639127573

This report instead seems to set the ratio more around 1:20000: https://www.statista.com/statistics/881541/bitcoin-energy-consumption-transaction-comparison-visa/

And this report 

I always take these estimates with a grain of salt because apple-to-apple comparisons are hard to make and the data is often incomplete.

But it's safe enough to say that the cost of a Bitcoin transaction is somewhere between 4 and 9 orders of magnitude higher than a Visa transaction (even a conservative estimate of 4 orders of magnitude more would be a cause of concern IMHO).

It actually makes sense that a BTC transaction is so much more expensive than a Visa/Mastercard transaction. The traditional circuits are optimized to process the highest number of transactions, while BTC is specifically designed to tightly control supply, number of processed blocks and frequency of processing. If you put such constraints, then your network must artificially introduce friction and inefficiencies. Sending money electronically after running a handshake between two trusted parties and exchanging a couple of messages is (and will always be) more efficient than sending the same money only after you have successfully brute-forced a SHA256 hash that meets some arbitrary conditions and synchronized your block with all the other nodes on the network.

Proof-of-Stake instead of Proof-of-Work is an obvious solution to the problem, since it replaces the energy-intensive hash cracking with the concepts of stakes and consensus, but it slowly pushes the pendulum back towards centralization. And, anyway, it'll never be really welcome in an industry made of cowboys who have invested a lot of money in big server farms, and who just want to keep mining rather than figuring out how else they can run their machines for a profit. 

 @7d199f28 I'm thinking more along the lines of the "traditional" transactions. At each point of interaction, power is involved. We're far from "all paper and coin" days.

I'm not putting crypto on a pedestal, but as someone who handled retail setups - there's a lot of power used behind the scenes. All people see is a register.

On a different level, I do like that old power resources are being updated and put back online. People think hydro has to be big. Not so. 

 @bca3d3cb Yes, there is definitely waste involved at every step in any economic system. But there's a big difference between the waste involved in the traditional paper-and-coin (or even "traditional digital") economy and that of the crypto economy (at least for the coins still based on PoW).

Outside of the crypto world, the "waste" is a byproduct of the economic exercise that creates friction and should be minimized.

An inefficient service that sucks up a lot of CPU and RAM thrown in the middle of the banking circuit would be seen as a problem. It would result in a slower throughput for the processed transactions, overloaded systems that may require frequent maintenance, loss of service, high bills from the server farm, etc.

In the PoW crypto world, instead, it is a *feature*.

Bitcoin, as of now, can't work without nodes that spend 100% of their CPU brute-forcing hashes. That's because it relies on a delicate balance of algorithmic throughput and supply control, distributed consensus and physical time required to crack numeric puzzles in order to work.

Waste and friction are required for the system to work, the same way that an electric stove requires a big fat and electrically inefficient resistive unit that sucks up a lot of energy in order to turn in into a lot of heat.

This difference between waste-as-waste and waste-as-a-feature is important to understand which models can scale and which ones can't.