There's a lot of energy to secure the network, but the actual energy usage per transaction itself is a small fraction of that misquoted number.
In a sane world, the energy use and price would be tied to the actual usefulness of the currency. They are not. In roughly 100 years the energy use will be set by transaction fees and the number of transactions, but currently it's only proportional to speculative will.
When transaction fees incentivize mining, the incentive to mine will certainly be much lower. Arguably more secure, too- the block reward creates a separate incentive for large players to dominate mining, which lowers the diversity of miners. It's not at all clear that the absolute energy consumption will be lower than it currently is.
[1]: https://digiconomist.net/bitcoin-energy-consumption/ [2]: https://ycharts.com/indicators/bitcoin_transactions_per_day [3]: https://www.coindesk.com/price/bitcoin
If you mean that the same amount of power would be used whether the blocks are full or empty, you are technically correct but in practice it's not relevant until blocks are consistently not full.
If transactions are critical for everyone, wouldn't that mean that keeping bitcoin mining alive is a fundamental part of transactions?
We could argue that fiat keeps databases running (if we ignore physical money) and bitcoin is keeping blocks mining active.
I don't think it is wrong to say megatwatts per transactions. We could ignore this completly and say 'the baseload of just keeping btc running is x megawatts per hour' and that would just ring the same bells.
I'm pondering if we could also say something like "btc itself as a cryptosystem motivates actors to consume megawatts per hour due to the interest in btc and the current fiat<>btc exchange value"?
Two glaring gaps in that concept...
First, energy use is not tied to transactions. An empty block uses the same amount as a full block.
And second, the transaction count we’re discussing is the count of settled base layer transactions. This doesn’t include the majority of transactions of value: those that occur off chain or through second layer transactions. An infinite amount of off chain and L2 transactions only need a single on chain transaction to settle.
Put another way - each validated block helps secure every single transaction that came before it.
I'm not going to argue that the energy consumption is worth it (I don't think it's sustainable myself). I'm only pointing out that energy per transaction is a bad metric if you want to have an honest debate. It's much more complicated than that.
Yes, I'm aware there's off chain solutions such as lighting but it's unclear to me whether those solutions are viable long term.
Transactions are the thing that matters. Simply securing the network achieves nothing if people cannot move money. That's why people measure the network with transactions/joule.
Instead, new people took it over and kept the throughput to the rate of a 56k modem. Now for the cost of a single transaction you can pay for enough hard drive space to hold the entire chain and enough bandwidth for a billion transactions.
This is like building a sidewalk instead of a freeway then saying you need to build a network of gondolas over it that just plunks groups of people down at different places on the sidewalk to move people.
Decentralised, fast, cheap: pick two.
It remains to be seen if there are good ways to do distributed currency, but there are certainly less bad ways.
If you want it decentralised it's going to be significantly slower and more expensive than centralised networks.
Increased cost is certainly not so straightforward. If you're sufficiently distributed, using already-existing hardware, with spare compute on extremely efficient devices, it could conceivably cost less. Even if a centralized server farm would be doing an order of magnitude less math, smaller processors use an order of magnitude less power to do that math.
Bitcoin proper has a central ledger- every miner needs to hear about your transaction to verify it. Lightning is a clumsy way of reducing how many actors need to be notified of your transaction. Better currencies include stuff like that as first class. It's all in the name of getting closer to a constant-number verification scheme that is closer to competitive with the O(1) of registering a transaction with a bank.
Centralized credit/debit cards exist so that a big wealthy firm can say yes, this person has enough money for this transaction and I will guarantee the transaction by paying for it even if they can't. What I would really like to see is distributed, automatic guarantees: when you buy something at a coffee shop, people running validators on wifi will pick it up and use their staked currency as insurance (hedged by the system) that they know accounts who trust this particular account, and the transaction is valid. More people staking and trusting this account means more trust by the larger system, which then only has to validate aggregated transactions. Anyone announcing themselves at a validator plugs in at a given level of aggregation, all of which have different staking/network/latency/storage requirements. Unlike off-chain transactions (eg lightning), the system is guaranteed at every level.
Any given transaction will still be validated dozens or hundreds of times. En bloc it will probably use tens or hundreds of times more energy than the server farms powering VISA. I'll be honest, I'm okay with that. I really like the idea of having a bank account that isn't tied to a company.