In other perspective, how average US household can use 867 kWh in a month is unbelievable, where let's say Germany is around 300 kWh.

To be fair the average "new" house is the US is almost twice as big as the average new house in the Germany.

Also there is a link between weather and energy usage. An average household in Louisiana or California uses way more electricity than New York's average.

Germany ranks very high in terms of renewable energy ratio and I think that's more remarkable.

For a full picture, however, you would need to count the total number of VISA transactions and compare that to the total number of Bitcoin transactions. I would guess the absolute energy foot print of VISA is still considerably larger.

But I am not trying to defend the energy usage of PoW, here. A single Bitcoin block can only fit about 3500 transactions and all miners of the world compete for being the first in solving the emerging Block‘s cryptographic puzzle. This is where the insane energy consumption comes from.

Other cryptos work differently in that respect and, hence, are way cheaper.

To put it in perspective, a single bitcoin transaction can power the average american's household electricity needs for... an entire month.

If you've looked at the transaction fees lately, you can definitely pay to power an American house for 1 month too (that is, assuming unreasonably that the house was powered at the same cost that most bitcoin miners pay for power: ~$0.04/kWh)

What is that in money, 50 bucks? (Live in Germany, we have very high energy prices, so I relly don't know. Single in 50sqm Appartment, paying 47euros per month for a 100% renewable plan.)

Depends where in the US -- In Texas (as long as there aren't any blizzard-driven blackouts with accompanying rate spikes), retail electricity costs something like $0.10/kWh. In California, it's closer to $0.20/kWh.

So roughly somewhere between $70-$150.

To put it in perspective, ~$1,000,000,000,000 USD is being secured by a decentralized system by using that much electricity. How much electricity/oil/other power is used by the US military to ensure the US government stays in power and the US currency retains its value? Surely orders of magnitude more energy than what is used for Bitcoin.

The number of transactions that are completed on the base layer is not the relevant metric. As the technology matures and as Bitcoin is more fully monetized, transactions will happen on higher layers. The "high" energy cost is necessary to secure the decentralized system.

Intuitively, it seems to me that the 741 kWh number for a single transaction cannot be a marginal number (i.e., the cost of the transaction itself without factoring in overhead). That's just way too high - it would be around $100 in electricity per transaction.

The difference is important because as Bitcoin adoption increases the number of transactions will increase and the cost-per-transaction will fall dramatically. But the actual marginal cost likely won't. (I bet the result would still dramatically favor Visa.)

No, that number sounds like a marginal number to me. Someone else commented that miners pay an average of like 4 cents per kWh, which works out to about $30 of electricity costs to miners per transaction. Recent per-transaction miner fees are currently in the $10-20 range, but don't forget the winning miner gets a reward that's currently ~$330,475 per block.

That's terrifying if true. 741 kWh per transaction seems absurd.

> The difference is important because as Bitcoin adoption increases the number of transactions will increase

No, the maximum number of transactions per block is fixed (actually, the maximum size of a block is fixed, but there's a limit to how small each transaction can be).

> and the cost-per-transaction will fall dramatically.

No, because of the limit on the number of transactions, the cost per transaction will increase (users have to pay a higher per-transaction fee to increase the chance of it being included in a block, and miners are incentivized to use the extra income from transaction fees to buy more power).

> No, the maximum number of transactions per block is fixed (actually, the maximum size of a block is fixed, but there's a limit to how small each transaction can be).

This has nothing to do with whether the number of transactions will increase.

>No, because of the limit on the number of transactions, the cost per transaction will increase (users have to pay a higher per-transaction fee to increase the chance of it being included in a block, and miners are incentivized to use the extra income from transaction fees to buy more power).

My point was that they seem to be counting overhead beyond the marginal transaction cost, and if they are, then the number they report will fall as the transactions increase. You seem to be talking about how the marginal cost per transaction will increase.

2750 Wh.gal. p.a. seems like a pretty high amount to me, so in other locales you could go s as linst twice as long

This is a dumb comparison. Bitcoin's energy consumption is necessary to secure a decentralized digital store of value. There is no alternative. It is electricity well spent.

Accordingly, you're using the wrong framework. Bitcoin's energy consumption prevents double spending, which is necessary for securing such a system. It isn't energy used to compute a transaction.

It's also a race to the bottom of maximizing spending/"security".