Hacker News new | past | comments | ask | show | jobs | submit login
[dupe] Bitcoin’s energy usage is huge – we can't afford to ignore it (theguardian.com)
83 points by subculture on Jan 18, 2018 | hide | past | favorite | 70 comments

This is another of the bitcoin related topic which has been discussed many times already in last couple of months:



I have also written a blog on this:


Isn't the root of the problem that fossil fuels are too cheap? If you tax carbon to include the costs of climate change then this problem will go away.

The idea is the switch to proof of stake will likely fix this particular problem if it works.

A lot of people seem to be ideologically opposed to proof of stake; they define "works" in a very narrow way that proof of stake can never satisfy.

Work as a concept is fine, but that massive amount of work having a direct correlation to massive amounts of energy consumption isn't ideal.

Is there a reason the proof of work can’t do something more useful, like protein folding?

It might be possible if the work was provably altruistic. Otherwise, any value the miner gains will be factored into the cost of mining.


Because of verification. It is easy to check whether the generated hash is correct or not. How will the verification for protein folding work?

It's easy to evaluate if the energy of one structural conformation is lower than another. (You don't need to determine at each step if the submitted conformation approximates the real protein at equilibrium, only that it's a lower-energy conformation than has been seen before.) But I'm not sure that is the only problem.

You want the proof of work to remain difficult for a PoW scheme. We have pretty good assurances that mindless cryptographic hashing is difficult; ASICs accelerate it considerably, but don't lower the Big O complexity of the problem. On the other hand, people keep searching for algorithmically faster ways to simulate protein folding, not just ways to hardware-accelerate existing algorithms. If one group of folders were to discover superior algorithms for searching the potential energy surface, that would be beneficial to science but potentially disastrous for a currency based on protein folding problems. For that matter, the overall incentives might make science worse off if ProteinCoin encouraged superior algorithms to be kept secret, for PoW advantage, instead of publishing advances in the open scientific literature.

Look into Primecoin (I have no stake) where the work done is to find the next largest prime number. This trudging work can be useful if we will it!

As I know, there is only 2 million Bitcoins unmined left. Once we mine all, aren’t we done with mining and miners ?

Miners are transaction processors, and instead of block rewards, they get paid in transaction fees. So, no, you're never done with them until you're done with Bitcoin.

No, because miners secure the blockchain. They will be rewarded primarily by transaction fees as the reward gradually phases out.

Sincere question: what's the energy usage for e.g. USD or EUR?

> Sincere question: what's the energy usage for e.g. USD or EUR?

Using VISA transactions for fiat currency, BTC uses 5K times more energy per transaction than VISA.


from the article you linked:

> Of course, VISA runs call centers, offices, and a whole lot else on electricity as well, which isn't counted in this comparison. But those hardly matter due to the extreme difference between the two figures.

--- I think this absolutely needs to be counted, because it is an operating cost of the entity. With bitcoin, there are no call centers because by design, the transactions are irreversable.

When pitted against the total cost of financial institutions, which all independently hold massive amounts of customer data, which should be considered a liability due to the recent issues with consumer data here in the West. This further increases the operating costs of each institution because they are running servers to house account data.

Imagine a global, catastrophic EMI event. Even a temporary one would cause introduce to these systems (like visa), a higher likelihood of race conditions where accounts will not reconcile, causing a higher cost to converge.

Bitcoin does not necessarily have this problem.

If we are truly worried about the future costs of running financial systems, maybe we should also add liabilities into the mix.

Visa could easily declare all transactions irreversible and close down their call centers if it made any economic sense to.

5k and growing by orders of magnitude, where it's likely the case that Visa is continuously reducing their energy usage.

VISA does not represent all (or even a significant percentage) or USD energy usage. You have all the banking and security infrastructure, employees, transport, physical manufacture and transport management, computing electricity. There is a lot more going on in the current monetary system than VISA.

Irrelevant question though, in many ways. There's simply no chance that bitcoin is replacing fiat money. I realize that offends the bitcoin True Believers, but fiat money is not going away unless nation states go away, and if that happens there's some more issues we have to deal with than the cost of a transaction.

That's even worse, because it means that Bitcoin's energy usage is additive on top of the energy used by the existing financial system.

No True Believers believe it is going away, they believe that it will diminish in purchasing power and trust (Relative to Bitcoin - or at least cryptocurrency). So far, this trend appears to be true and continuing. Who knows how far it goes? Of course you will always be able to pay taxes in USD.

Whether it's relevant or not wasn't my concern, as I wasn't even trying to propose that any cryptocurrency replace fiat. I was simply curious.

It's zero. Exchanging banknotes or coins for goods or services does not consume electricity. Can't believe this was "sincere question".

Coindesk did a write up of this.


Looks like data was taken from places like: https://www.ecb.europa.eu/stats/policy_and_exchange_rates/ba...

and rest of the numbers were taking into account of how much each note costs to make.

"The Euro publishes sustainability statistics on their currency, and according to latest estimates, 3 billion banknotes printed in 2003 had an equivalent energy impact of 460,000 60W bulbs switched on for a year, which equates to 240 million kWh, or 0.87 million GJ. With circulation now at 15.8 billion notes, this would scale up to 4.6 million GJ (European Central Bank, 2007). To get to a global figure, for the purposes of this report, I will be multiplying this figure by a factor of four (i.e. a proportional share of global M0/M1 money supply). Therefore, we reach a figure of 18.4 million GJ, which would correspond to almost 3.07 million tonnes of CO2 equivalent."

This is just making a bank note, nothing about cost to distribute (guarding money, banking system, credit system).

According to https://digiconomist.net/bitcoin-energy-consumption, Bitcoin is currently at 42 TWh. That's over 20 million tonnes of CO2.

If we assume energy costs to produce annual supply of new banknotes and coins in entire world is at 1 TWh. You see how insane energy consumption of Bitcoin really is.

And that's comparing an entire money supply of the world to single cryptocurrency.

Banks cannot be part of the equation as the need for banks would not cease if the world would switch to Bitcoin or similar.

Fair point. All I was trying to show was that currencies like the Euro or USD has greater than zero energy usage to create/distribute.

I don't think BTC will stay the top crypto if it cannot resolve its current issues (though new ones will come after). Articles like takes stance, whether they know it or not, of this is here to stay for a long time and we need to deal with the ramifications. If that is true, more people are going to work on solving this issue either, or all of, through energy regulation, better hardware, better renewable energy, mass adoption of different coins, or moving to POS (or anything other proof system) instead of POW.

> Fair point. All I was trying to show was that currencies like the Euro or USD has greater than zero energy usage to create/distribute.

How many cash transactions occur in the world in 1 year? The number must be over 7 trillions assuming average person on this planet makes 1,000 cash transactions per year.

Now, let's divide 1 TWh by 7 trillions, we will get 0.14 Wh per cash transaction which is equivalent of having 60W light bulb switched on for about 8 seconds.

Considering one Bitcoin transaction currently consumes 345 kWh of electricity. That's equivalent of having 60W light bulb switched on for about 8 months.

So yeah, technically you are right. 8 seconds is not zero. But it's hell of a less than 8 months.

I think it's a fair question; each transaction of paper money causes wear so it might make sense to amortize the cross of minting over each transaction. Plus transactions between banks would use some energy.

One can only assume the energy cost is negligible per transaction compared to bitcoin though, simply because if you scale up the transaction volume of bitcoin to what a real economy would need it's more than we produce.

Not to mention adding all of the overhead for legislation, note and coin design, sorting, transporting, anti-counterfeiting...

Many bitcoin proponents believe that bitcoin can fully replace the current financial system so whenever the energy consumption problem comes up you should expect to hear someone talking about how much banks spend on air conditioning every year.

Of course banks wouldn't go away with Bitcoin. There are still physical branches despite the ubiquity of online banking and credit/debit cards. People like having humans to interact with. And Bitcoin is trying to add in the Lightning network, which requires large trusted entities where people can store their coins for off-chain transactions, IE banks.

I doubt that they still believe that, as much as they are highly incentivized to claim that. Bitcoin is a disasterous enterprise as it continues to scale poorly. But hey, some people get rich (for a while) and others get to have a waking ideological wet dream, so who cares if it’s a tragic exercise in “printing” entropy and calling it value.

Whataboutism isn’t going to change that.

Besides, hasn’t the world suffered enough for the sake of grasping ideologues? Does everything need to be seen through these same bland lenses? This is a truly maddening and mad use of a currently limited resource with globally damaging externalities, and it produces nothing except waste heat.

If the average Joe really understood this, and just how modest the real potential of blockchain tech probably is, there would be lynchings.

It depends on how far you want to go.

The cost per transaction would be interesting in real terms.

What's the comparison between:

- Manufacturing physical currency / Manufacturing the hardware to mine coins - Cost to run POS systems in retail outlets / Cost to run bitcoin network - Cost to run global fiat exchanges (i.e. wallstreet) / Cost to run global bitcoin network

Obviously you could drill down as far as you wanted. I'd imagine it would be quite difficult to draw up such a detailed comparison though.

This is anecdotal but my family operated a store in a rural vacation area and over half of our customers used cash. In the 2000s, we'd have to drive the cash 30+ minutes to the bank, twice weekly to mitigate the risk of a robbery. We sold the business, but more electronic transfers hopefully reduces the required driving miles, and overall cost to transact in cash.

I can't find the study right now, but recently saw data suggesting that the cost of money handling is still several times cheaper on a per dollar basis than electronic transactions.

If a business accepts any cash at all, the marginal cost of another cash transaction is really low. By your own antidote, your store still saw the value of accepting cash transactions, yeah?

are you sure?

all of these surely have a substantial eco footprint - money production, for both paper and coin - armored cars picking up cash and bringing to the bank - physical locations, banks, check cashing places

That's pretty much the only time it clearly consumes energy; and even worse, manpower.

The lights at the bank of america HQ don't keep themselves on.

Financial transactions are more than just exchanging pieces of paper in person.

what about running bank of america and visa?

If you're suggesting virtual currencies won't need trustworthy managers, I think you better check your wallet.

This article is based on some incredibly bad estimates, and reaches an outrageously large conclusion.

Let us assume that 100% of all mining rewards go to electricity (and none to buying mining equipment, paying employees, servers, or networking), and let us further assume the electricity costs $0.01 / KWhr, which is cheaper than one can buy basically anywhere.

From flippening.watch, BTC paid miners $20,797,200 yesterday, for 2,079,720,000 KWHrs. This translates to approximately 87 GW of electricity, 24/7. 87 GW is approximately the output of the 21 Palo Verde nuclear plants in Arizona, which according to the US Energy Information Administration, can output up to 3937 MW. https://www.eia.gov/tools/faqs/faq.php?id=104&t=3

In other words, Bitcoin is a drop in the bucket. Let's put this myth of massive energy usage to rest, and at the cost of being too snarky, and let us ask our journalists to learn basic math.

edit - I made a math mistake, which I have corrected. The overall point, while I was off by several orders of magnitude, however, holds.

edit 2 - I retract the entire comment. It appears the article is true, and that, ironically, I made the same mistake I accused the journalist of. I'm going to leave the comment up for posterity, but I retract it.

Your original point does not hold after correcting your 3-orders-of-magnitude error. The Guardian article estimates that bitcoin is on track to use 42 TWh of electricity this year. Your corrected estimate, 87 GW, is 762 TWh annually.

Your contention that the Guardian article made an "outrageously large" estimate of bitcoin electricity consumption is contradicted by your own estimate, after you fixed the units-conversion mistake in it.

How did you go from 2,079,720,000 KWHrs to 24.07 MW of electricity, 24/7?

My calculator and Google (search for "2079720000 kWh per 24 hours to MW") says it's more than 86,000 MW (so ~86 GW).

edit: corrected units from TW to GW, as pointed out by philipkglass

86 GW, not 86 TW. 86 TW would be several times world primary energy consumption.

You're right, thanks! It was an unfortunate typo.

How are you getting that 2,079,720,000 KWHrs in 24 hours is 24.07 MW?

EDIT: This is too perfect. 1053r calls out journalists for making a basic math mistake. In doing so, he makes a basic math mistake. I ask him about it, then get downvoted by HN.

2,079,720,000 / 3600 / 24 = 24070.8333 KW, or 24.07 MW.

Why are you dividing by 3600? The numerator is already in kWh units (kW hours), it's not kW seconds.

Yup! Sorry! Will attempt to edit original. Should have said 86GW.

It's a problem because it's a digital currency? Because many other things that could be considered "surpeflous" also consumer massive amount of electricity. Or for example (not superfluous but potentially equivalent) how much electricity all the global banking system uses? I bet that it's more than "a nation".

fast forward to a Dyson sphere for crypto mining

How does a nation expand credit with a cryptocurrency? Like as in the case with fractional reserve banking?

Unfortunately it has all the disadvantages of a hold based system. It also has the advantages, but imho those are essentially nonexistent in any practical sense (yes, inunderstand there is a political argument otherwise).

In the hope avoid down voting, in case you don’t know the problem of gold standards are three serious ones. First, it’s deflationary as the gold supply doesn’t change as economic growth occurs. Second, it’s decoupled from economic activity (e.g. you increase the money supply every time you use a credit card and decrease it when you pay off your bill — and that’s only one example). Third, it can suffer inflationary crises (yes, despite being deflationary!) when there’s a big gold strike (as happened with the California, Yukon, and Victorian strikes, among many others).

The argument FOR it is that governments can just print money if they get into trouble and inflate the currency. This usually screws the people doing the inflating as well (see: Venezuela) but can be pretty dreadful (e.g. Weimar). But the problems above are also terrible.

A nation-state (or bank, or a credit company, or an individual) could simply create fiat crypto, the same way they create banknotes, bonds, or IOU's - out of thin air.

I.e.: they take gold or similar valuables, and give you a contract for it, authenticated by a blockchain.

They can expand this 'money' supply by lending: They create new coins and give them to you _without_ getting your gold right away.

In a sense they lend money they don't have. This is basically how fractional reserves operate; everyone hopes that not everyone tries to get the gold back at the same time. And if you think that sounds dodgy, well that's just how banks started out, before paper money was invented, only without the technology to prevent forgery!

Unlike IOU's and other form's of fiat, some sort of crypto should be immune to forgery, if implemented properly. It's just a tradable contract, backed by a blockchain.

However, it still depends on trust and/or legal enforcement (hopefully the later) that the gold will be there to paid out on demand. Like any other fractional reserve banking system, enough bad-debts can cripple it, and it is not immune to collapse if there is ever a 'run' on it.

The next trick (once everyone is accustomed to exchanging crypto) is to have everyone accept that instead of getting their gold back (ever), you'll allow people to use the crypto to offset their tax liability.

Voila - money for nothing!

Here are some thoughts on how it could work. Not completely thought out, but something:

Before we had electronic bank records, banks could print "bank notes" which were as good as money if you trusted the bank. Now that we have electronic records they basically "create" USD by adding a deposit to your account while only having ~10% of that actually stored away in dollars.

If your account was truly in BTC they obviously couldn't just increase your balance, since it's a cryptocurrency and they can't just make something up.

One option would be for banks to have their own tokens, similar to bank notes of old. All banks would accept the tokens of other banks 1:1 and convert them to BTC on demand, but it would allow expansion of the "money supply" like we have now with dollars.

> One option would be for banks to have their own tokens, similar to bank notes of old. All banks would accept the tokens of other banks 1:1 and convert them to BTC on demand, but it would allow expansion of the "money supply" like we have now with dollars.

Why not create one big bank that can issue its own tokens that any bank will accept? We can call it a "central bank" or something. And then maybe after many years we can decouple the value of these central tokens from the value of bitcoin, instead allowing the central bank to control the total supply of tokens based on some kind of "monetary policy" to keep prices stable.

Sounds good. The future of crytocurrency is bright indeed!

Any online site that sells bitcoins for fiat, like Coinbase, could just keep the fiat (or use it for anything other than obtaining bitcoins) and increment your bitcoin balance. Their total bitcoin holding may be less than the sum of their users’ balances, and no one would notice unlrsss there was a “run” of users attempting to withdraw bitcoins into their own real wallets. We’d probably all consider that blatant fraud.

The same way we do with cash. When you put your cryptos in a bank they will lend out some portion of it and in return you'll be paid out some interest rate. The incentive to storing your crypto in a bank will be that you receive that interest rate, potentially that its insured, and that its secured. Perhaps the interest rates will need to be higher to encourage someone to deposit their cryptocurrency in a bank, its hard to say.

it doesn't.

High electricity consumption is the trade bitcoin made in its design for a stable money supply, it's not an inherent problem with cryptocurrency.

It's the tradeoff made for a distributed money supply. Satoshi just as easily could have made a fixed supply of money which is also stable and doesn't rely on proof of work.

That said, at the time, I don't think this was really the tradeoff being made. When bitcoin originally came out, proof of work was the only decentralized consensus mechanism available.

Having a distributed money supply makes it unstable, since you've tied an incentive to creating bitcoin.

Money supply would then be equal to how many miners there are without adding a difficulty, and the amount of bitcoins would go through the roof. Difficulty was clearly added as a solution after this to keep the money supply stable.

Note I didn't say it's necessary to waste electricity to have a stable money supply, only that the reason Bitcoin does is because it keeps it's money supply stable.

Difficulty adjustment is used to keep block times stable. Without this adjustment block times would continue to go down until you would eventually end up with a chain that has much more frequent forks of greater lengths. You eventually end up with a state where the current head becomes indeterminate and all sorts of security issues become possible, including double spends.

Bitcoin could have easily based block rewards on timestamps to keep the supply of money consistent, similar to how the current difficulty is calculated.

I posted on FB while working on my degree and got a 5 year memory about it recently. I visualized the miners turning the polar bears black, and I was a good liberal who turned off my lights. Ethics would win. "It was a cool prototype" that I forgot about. Every CS student can figure this out, so it's surprising that we've let it go. Incentives, right.

Polar bears have black skin with white fur

My car has steel panels with white paint. It's still a white car.

What a ridiculous article. Household air-conditioning wastes energy. Running your television wastes energy. Serving up webpages for The Guardian is apparently a waste of energy. If the problem is that the price of energy does not encompass the negative externalities it creates, then we need a Pigouvian tax on it. The price rises, and miners reassess whether or not they are profitable in a carbon-neutral world. The author seems to think that market participants using energy in a profitable manner are at fault here, rather than the policymakers who are not interested in tackling the Greatest Moral Challenge of Our Time (TM). But sure, jump on the bandwagon and write op-eds while Bitcoin is in the news.

> What a ridiculous article. Household air-conditioning wastes energy. Running your television wastes energy. Serving up webpages for The Guardian is apparently a waste of energy.

All of these provide some form of value to the user. I have yet to see what value bitcoin provides over USD. Okay, so we now have another medium of exchange, so what? Why is that worth the 1000x energy consumption per transaction over payment processors like Visa?

All your examples, by the way, have gotten more and more energy efficient over the last 50 years. Bitcoin is going in the opposite direction for payment processing.

Thank you for actually engaging! Let me preface this by saying I'm usually the first to deride Bitcoin as a payments platform. However, using energy to mine Bitcoin does provide value to the user - profits over and above what the power bill and ASICs/GPUs cost. And you're right - for a payments platform, it is a regression in terms of energy consumption.

That, however, does not change the fundamental issue - the severe underpricing of emissions-heavy power generation. The externalities that the author is so worried about (in the form of climate change) are not priced into the going rate for electricity. Blaming bitcoin miners is silly - if I pay for my power, I should be able to use it however I see fit. The alternative leads to absurdity, such as a law that states your aircon may be set no lower than 22C in summer, as anything more would be "a waste". I pay for it, I use it, bugger off.

The external social cost of electricity generation needs to be internalised into the private end-cost - a carbon tax or an ETS. Pick one, follow through, and stop blaming rational agents for acting rationally.

EDIT: Spelling

Applications are open for YC Winter 2022

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact