Comment by nonbel | Hacker News Reader

Comment by nonbel | Hacker News Reader

I posted this on a similar topic:

  # kilowatt-hours per transaction
  # https://motherboard.vice.com/en_us/article/ywbbpm/bitcoin-mining-electricity-consumption-ethereum-energy-climate-change
  btc = 215 

  # Transactions per day
  tpd = 3e5

  # convert btu to terawatt-hours
  # http://www.dvirc.org/how-much-energy-does-an-office-building-consume/
  officeEnergyPerYear = 1.4e9*.293071/10^12

  # https://www.usatoday.com/story/money/business/2014/10/05/24-7-wall-st-banks-with-most-branches/16648133/
  nBankBranchUS = 94725

  # Convert to kw-h to tw-h
  btcEnergyPerYear = btc*tpd*365/10^9

  >   btcEnergyPerYear/(officeEnergyPerYear*nBankBranchUS)
  [1] 0.6057412

I have no idea if the source numbers are correct, but this is a better statistic. Mining uses the same amount of energy as used for ~58k average office buildings. So the worldwide bitcoin mining energy consumption is about half of that currently used for banks only in the US. This is assuming banks are "average" (probably higher due to security concerns) and of course that banking doesn't consume energy in any other way.

https://news.ycombinator.com/item?id=15614966

-----END-----

Has anyone actually tried doing one of these comparisons yet? I mean going beyond just back of the napkin? Can we get estimates so about whether credit cards consume more electricity than many countries? Without that this type of article falls flat. Actually it is interesting that it is so easy to estimate the total energy consumption in this case.

Using the same amount of energy as nearly 60,000 office buildings to perform ~300,000 transactions per day seems catastrophically bad. That's about 5 transactions per day per office-building-equivalent.

MasterCard, for example, averages something like 2400 transactions per second. That's around 600 times more than Bitcoin. I don't know how many office-building-equivalents MasterCard uses, but I doubt it's anything like the 60 million that would be implied by equivalent efficiency to Bitcoin.

The US banking system as a whole does something like 4 million transactions per second. Using half the energy that these banks use in order to do 3 transactions per second doesn't sound good at all. Even if your estimate for bank energy usage were 1000 times too low, it would still be ridiculous.

Thanks, another interesting number would be number of people employed in each industry. I found for US banks it is about 2 million: http://www.usbanklocations.com/bank-rank/number-of-employees...

I have no estimate for bitcoin in the US (I'd guess 1k - 100k?). Also at the end you slip into comparing worldwide vs US numbers. So both energy and transactions/s will be higher for the correct comparison.

It's not appropriate to look at the number of people working in the finance industry, as the vast, vast majority of finance industry is not employed in transactions; almost all the headcount goes to handle lending, credit, risks and investments. A few percent seems appropriate from what I recall in the banks I've seen, so <100k people maybe for all the payments in US.

Only a tiny fraction of the power goes into performing transactions; the total power consumed is independent from transaction rate. Bitcoin in 2009 used vastly less power than today yet had the same transaction capacity. The power consumption scales with the price of Bitcoin[1], which is determined by demand and most of the demand right now is driven by future price speculation.

Interestingly, the new cash-settled CME Bitcoin futures contracts, due to start trading in December, will allow speculation on the future price to take place without requiring possession of actual Bitcoin. These futures may both increase Bitcoin adoption and reduce the upward pressure of speculation on the price.

[1] More accurately, the value of power consumed for mining Bitcoin approaches the value of the block reward plus transaction fees.

You could say only a tiny amount of the power is used for transactions. But given that the transactions are the entire point of the network, all power used by the network ought to be counted towards the transactions, just like we’d count the power used by the lights in a bank towards their transactions even though the lights don’t directly do transactions.

The fact that the power used by Bitcoin is essentially independent of the number of transactions is interesting, but it’s not a plus for Bitcoin, since the network is at capacity already.

This assumes that the value produced by bitcoin transactions is equivalent to the value produced by banks operating, though, which seems a rather huge assumption? The stuff banks do on a daily basis is very, very important, and they do a lot of it.

Like, a trivial example: Let's say I can eventually pay my rent with bitcoin instead of a bank wire. I'd have to do that off-chain or on an exchange or something, right? What happens once you factor the cost of operating those exchanges in? My landlord probably needs to convert that BTC to USD at some point, so then banks and wires are involved to transact the USD.

It's trivially the case that the value produced by bitcoin mining exceeds the cost of the electricity, else nobody would bother.

#cough# JavaScript cryptocurrency miners in adverts, #cough# cloud computing with stolen resources, #cough# anticipation of future bitcoin value when estimating cost effectiveness

It's trivial that the value exceeds the price of the electricity. Unfortunately, electricity is often priced well below cost once you account for externalities like pollution.

Aren't bitcoin transactions dependent on miners continually mining in order to verify transactions? It sounds like bitcoin is relying on a perpetual profit margin for mining, in order to make transactions possible.

This is the point of transaction fees in bitcoin. Fees get 'attached' to the block that is mined and go the miner. So even when all bitcoins have been distributed there will remain a profit motive for mining. The difficulty in mining is also automatically algorithmically adjusted, so if the industrial scale miners move onto other things - suddenly it might become profitable to go back to GPU mining. And the fewer people mining, the more profitable it would become.

Is it? And are those who are consuming the electricity paying the entire cost for it, including impacts to the environment?

Surely there is also a lot that banks do that isn’t important, or that is even harmful to society.

Bitcoin also does things harmful to society -- like enabling money laundering and tax evasion.

Reply on news.ycombinator.com