I provided this statement to some journalists who contacted me to get another viewpoint on Alex de Vries's paper:
His findings are not completely accurate, but also not completely wrong either. My own formal well-documented estimate is that Bitcoin used 0.09% of the world's electricity as of January 11th, 2018 (http://blog.zorinaq.com/bitcoin-electricity-consumption/). After 4 months of growth I estimate we are around 0.15-0.20% today. Therefore reaching 0.50% by the end of 2018 is plausible; not likely but plausible. That said, overall I think Alex de Vries's figures are rather inflated because he makes two errors. Firstly, his model assumes that manufacturers sell at cost and make zero profits, which allows miners to spend more on electricity. In reality we've seen Bitmain, the largest manufacturer, make billions of dollars over the last few years. No one expects them to make zero profits by the end of 2018. Secondly, de Vries makes the mistake of assuming that all cryptocurrency mining chips produced by TSMC on behalf of Bitmain end up in Bitcoin mining machines. In reality many chips are fabricated for cryptocurrencies other than Bitcoin. This altcoin sector is rapidly growing. It wouldn't be unreasonable to assume that Bitcoin mining machines will represent less than half of Bitmain's sales by the end of 2018.
> After 4 months of growth I estimate we are around 0.15-0.20% today. Therefore reaching 0.50% by the end of 2018 is plausible; not likely but plausible
What model are your forecasts based on? From what I could tell, the featured paper’s model allows us to forecast future consumption but your research only presents a current-day lower bound (Jan 2018 in this case). I may be wrong though
The deVries estimate of ~2.55GW/26PH/s (mid Mar 2018) tracks closely with mrb’s lower bound estimate of 1.62GW/16.2PH/s (mid Jan 2018). In just two months the power consumption and hash rate has risen 60%. Two months later, mid-May 2018, the hash rate is around 31 PH/s, an increase of 20%.
This is on par with historical increases documented by mrb (325 TH/s in late Feb 2017 to 640 TH/s in late Jul 2017, an increase of 100% in 5 months). If we see an average growth of 100% each 4 months, then we will easily surpass 100 PH/s by year-end, which is around 10GW, or around 0.4% of the world energy production.
Thanks for the response and analysis. I think I have qualms with how mrb exaggerates the difference in their conclusions. He says Alex' numbers are "rather inflated" and has told me before that Alex makes "fundamentally flawed assumptions" and has "holes and flaws" in his logic, when really when you look at the difference in their forecast it's 0.5% versus 0.4%... both numbers support the general idea that assuming the technology brings no efficiency to society, Bitcoin is an environmental catastrophe.
Alex's BECI model claims Bitcoin currently consumes 7.6 GW. Whereas this paper is more reasonable and states the future consumption might grow to 7.6 GW by the end of 2018. These are 2 very different statemetns. What I call "fundamentally flawed" is the BECI claim, not this paper.
My model actually provides both an upper bound and a lower bound current-day figure (explained in http://blog.zorinaq.com/bitcoin-electricity-consumption/) 0.50% is not predicted by my model, but it's just a target that seems plausible if the current growth stays the same.
For the first point, the article explicitly examines at Bitmain’s profit margins on their devices, where they find e.g. a profit margin of 56% on the Antminer S9. Although Bitmain might be headed for increased profits in the short run, from an economic standpoint it is reasonably to assume that market participants will converge towards zero economic profit - this is a standard economic assumption.
For the second point, he’s fairly careful to specify that these are estimates from a third party (Morgan Stanley) and does not explicitly specify that he expects all of them to be Bitcoin mining rigs. The estimate of power consumption he produces here, 8.92 GW, is used merely as a check on his main figure of 7.67 GW obtained through the economic analysis; he does not directly rely on the Bitmain production figure. Plus, he also notes that this 8.92 figure doesn’t take into account the ~30% of the market not owned by Bitmain.
Yes, but given this convergence hasn't happened in years, it is unreasonable for de Vries to suggest it might happen in 7 months.
For the second point, he definitely assumes 100% of the production ends up in Bitcoin mining machine. His source says "cryptocurrency" chips, but de Vries assumes "Bitcoin" chips.
I'm curious about this actually: what is the power consumption of the whole ensemble Bitcoin+altcoins? From what you're saying + the original thread, it looks like almost sure that electricity used by those combined would be at least 0,5% and could even be a higher number, no? Even if Bitcoin is the biggest one, the others are quite a lot which makes them not completely discardable I guess.
PoW isn't "Proof of Work", it's "proof of money spent". The "work" is only here to make crypto look sophisticated. We could replace "work" with videos of miners burning stacks of cash, & crypto would work just the same, without killing polar bears. Downside: it would look stupid.
Cryptocurrency and Blockchain is easily the most misunderstood technology thats existed. I blame ICO youtube videos that propagated the myth that Blockchain is going to cause a revolution.
Blockchain ONLY exists to decentralize and validate.
This is useful for currencies, voting, timestamps, and maybe a few other niches.
Centralized applications are cheaper and faster than blockchain. You wouldnt use blockchain unless you needed the decentralized validation worth paying for.
Users like OP mentioned doesnt understand what the Work is doing. Most people who 'know' crypto, have no idea what blockchain is.
The only people worth listening to are ones that spend crypto and have developed. Everyone else are parroting myth.
That doesn’t make any sense. The cost of a particular PoW can vary enormously. Mining a Bitcoin block in Hawaii using a CPU will cost orders of magnitude more than doing it in Iceland with an ASIC, but they are treated identically in the network. It’s proof of work. It just so happens that work costs money.
It should be Proof or Waste, or Proof of Heat/Entropy. The work being done is difficult, yet trivial and only valuable because it yields coins. Miners convert free energy into heat, for funny money, and it doesn’t scale well. Sure, maybe PoS could work (I doubt it) and maybe renewables could be the future of mining. Unfortunately we’re stuck with a majority of miners burning coal in China to show off the required volume of entropy to cash in.
When this fad dies, it will have been historically significant in terms of pure waste.
Similarly, my joke is that bitcoin is an alien species or hostile AI's bloodless, costless, infinitely scalable weapon to get potential spacefaring competitors to abort themselves by subverting their star's energy to fruitless labor.
The government already mostly uses accounting instead of physical money so moving to a really expensive way of doing accounting probably wouldn't save money.
I think they were trying to intimate that a large-scale adopted cryptocurrency could replace large swaths of financial infrastructure (this is a sort of pie-in-the-sky ideal, but something like it could happen) - picture a cryptocurrency replacing the federal reserve, VISA, the european central bank, brick-and-mortar bank branches and credit unions (partially), etc. It couldn't replace everything, and new things would become necessary, but I think there is potential (even if small) for it to provide enormous value at least on the order of its energy consumption.
Traditional ways of doing all those things are faster and cheaper than blockchain.
Transaction clearing between banks isn't particularly fast, but that's because of legacy legal issues more than it is anything technological. The actual transfer of large sums can happen very quickly.
"If everyone agreed with me, there'd be no disagreement!"
Most of the things that crypto proponents tout as features (transaction irreverisibility, lack of inflation, etc.) are in fact bugs, and they just don't realize it because they don't realize how finance and economics actually work.
Tapping your watch and wondering "why isn't everyone using crypto already??" is never going to yield useful results, so you do actually need to address these issues.
>Most of the things that crypto proponents tout as features (transaction irreverisibility, lack of inflation, etc.) are in fact bugs, and they just don't realize it because they don't realize how finance and economics actually work.
These are not inherent features of cryptocurrencies, there are cryptos with reversible transactions and inflation.
> (transaction irreverisibility, lack of inflation, etc.) are in fact bugs
They weren't bugs for the thousands of years when gold was used.
> they don't realize how finance and economics actually work.
You probably mean the last 40 years of finance and economics, since our modern monetary system has really only been around since Nixon took the US off the gold standard after the US over spent due to the Vietnam war and Charles de Gaulle started to exchange his dollars for the quoted rates of gold.
Transactions with physical currency are reversible, just not easily so. And solid gold was rarely used as a day to day currency, so you would have an amalgam which was subject to inflation as the government either changed how much gold backed it, or phsycially reduced the amount of gold in the coin.
For thousands of years people used actual gold, especially countries moving money back and forth between them.
Inflation in its current form didn't really exist. The exchange rate for gold of a country's currency didn't change much, if at all. The dollar went from $20 per ounce, to $35 per ounce during FDR, then lost the vast majority of its value, getting us to over $1,300 per ounce that we have now.
Physical money requires the person giving it back physically. Crypto currencies have a chain of transactions, you can just send the balance back to one of the addresses it was sent from. There isn't much difference there.
Isn't there an argument that a lot of the mining is done in places where energy is cheap because it is in abundance and would be unused otherwise? Near hydroelectric sources?
This is absolutely false. A fairy tale told by those who want you to buy their crypto currencies. Take Wenatchee, a village in Eastern Washington, USA, for example. Wenatchee has very cheap power because the water dam is very close nearby.
When crypto currency miners poured into the city, the power grid got overloaded and the local authorities had to raise prices drastically. The city finally decided to ban all mining but there are still many rogue miners left, operating from basements around the city. They let the local population foot their costs by mining for cheap and leave when prices need to rise to support the grid.
Mining crypto currencies is morally corrupt and hurts our environment more than anything ever before without generating any value.
Stop being deliberately dense. The act of "moving ions" has huge externalities, mainly in the form of pollution, and that is obviously the morality that is being referred to here.
Your comment is like saying you're baffled that shooting someone is immoral because it's just "moving metal around".
Every zero-sum adversarial system would be a lot more efficient if they accepted the judgement of a third party instead. How much energy could we save if we dismantled the court system, elections, and pro sports?
I think DAG coins could be a lot more efficient, but I'm not totally solid intellectually on how they achieve consensus on conflicting spends so I'm not sure it will work as a general solution. The only one I've used, nano, has I believe no mining reward, and I think consensus is achieved based on a voting process weighted by proof-of-stake among online nodes -- the incentive for participating in the system is that the system has value for you in other ways. I'm not closely following ETH PoS research, but it will be interesting to see what they can accomplish, it could create a huge reduction in energy consumption.
> Every zero-sum adversarial system would be a lot more efficient if they accepted the judgement of a third party instead. How much energy could we save if we dismantled the court system, elections, and pro sports?
None, since players would devote their energy to influencing or becoming the third party instead of playing these games.
DAG coins just distribute the mining work across the entire network including edge nodes. That's not necessarily more efficient but it is more decentralized. Of course it remains to be seen if DAG systems can be robust against attack. I'm familiar with several attacks against DAG-type cryptocurrencies that don't seem to have really good answers.
If this is correct then if we had a cryptographic lottery players would devote the same amount of "work" to attempting to game the lottery that they presently burn by pumping electricity into mining ASICs.
The point of cryptographic lottery is that it's not proof of work scheme where dong work can improve changes and be profitable.
It's possible to define a decentralized objective consensus without PoW and make it unprofitable to use sybil attacks and other methods to game the system.
If you could do that so easily, you'd already do it in aluminum plants, which are kinda like bitcoin already -- electricity in, aluminum out. Suffice to say, it doesn't really happen.
40% of gold consumed is in investments, 50% in jewelry, and 10% in industry, so I'm not totally sure how it works out in terms of energy per unit of gold, but it's likely some mining of gold would be necessary even if you stopped using gold for jewelry and investing. But you could get rid of bitcoin (and other crypto currencies) without any real impact.
perhaps. you could get rid of most of things in our society without much impact, since we basically need food, water and shelter. but lots of things are nice to have, so that's kinda irrelevant
I think cryptocurrencies have very few practical applications at the moment. So getting rid of them would have much less effect than getting rid of gold for example.
yes, that's because they are 10 years old as opposed to 10,000, but I am not sure where you are going with this. At the very minimum, getting rid of them would cause investor losses of hundreds of billions.
gold mining uses a ton of energy, more than bitcoin mining. To the extent that people substitute gold for bitcoin in financial portfolios, this is positive since net energy waste is lower.
It seems like a bad comparison, since you don’t need to mine gold just to move it around. Transferring ownership can be accomplished for the amount of energy it takes to send a small electronic message.
From some quick searching around, volume on the exchanges is around 500,000/day while volume on the blockchain is over a million. If that’s accurate (and it may not be, I’m not an expert on this and the sources I found might suck) then most volume is on the blockchain by a pretty good margin.
I don’t know how much gold is actually shifted around physically, but the trading volume in a year is several times the actual amount of gold that has been mined, so I suspect the proportion of that which involves physical movement is tiny.
Hm, how much energy DOES it take (how would I compare? not sure) to produce equivalent number of dollars? Is there maintenance cost? Interesting line of thought to me.
Most dollars are virtual, so the energy cost is just sending some small messages around. It’s orders of magnitude less than needed for a bitcoin transaction.
Physical currency probably takes more energy, but printing is an old and cheap technology, so I doubt it’s comparable.
A thought experiment: how much energy should Bitcoin use?
Bitcoin's mkt cap is ~$140B [0]. The world's GDP is ~$127T [1]. If we assume:
1) BTC market cap is representative of its economic output in accounting terms, or is at least an upper limit
2) Energy use by product should be proportional to its output
Then at 0.11% of world output, Bitcoin is at least ~5x more energy intensive per unit of value than the average product.
(This is obviously not wholly accurate. For one, market cap != annual value. If accounted for, that might make Bitcoin several orders of magnitude less efficient. And assumption 2 is probably a linear approximation to a highly nonlinear relationship. But I propose this as a fun thought experiment that questions the energy-value relationship.)
Totally agreed. For the thought experiment, make the super liberal assumption that most value was produced and realized in the past year (or a single year), and it becomes comparable. Otherwise, the inefficiency is a hard lower limit. So in the extreme case, the lower bound is 5x inefficiency. If we assumed a value equivalent to 1-5% of market cap is realized every year, it becomes 100x to 500x inefficient.
>A thought experiment: how much energy should Bitcoin use?
If it used proof-of-stake, then it should be a negligible amount compared to what it uses now.
(Disclaimer: proof-of-stake is still being researched. Many altcoins presently claiming to be proof-of-stake rely on centralization (peercoin), likely have issues, and/or are underverified. I'm cautiously optimistic about the Ethereum community's thorough research into proof-of-stake.)
I have hypothesized that long-term, crypto currency will be made illegal to some degree eventually for one of two reasons or both:
1. Governments use fiat currency as a lever for it's monetary policy and cool/heat the economy and job market. If crypto use becomes widespread enough to weaken this lever, I could see pushback from governments.
2. This, environmental concerns. Crypto has to be the least-green modern technology by ALOT. Once the public catches on, I think there will be significant push back. This is of course largely dependent on the population deciding to side with the environment, which history has shown us has to be a mixed bag.
More likely is the application of restrictive regulation in order to combat illegal activity that involves bitcoin. Money laundering, kidnapping, extortion, and bribery are kind of a big deal.
Bitcoin should be made illegal. I don't want the ability to anonymously and irreversibly pay someone. If I have this ability, then bad actors have an incentive to try to force me to do so. We've seen this in the wild already - both kidnappers and computer virus writers have included "send BTC to this address" as a demand.
Not every ability and option is positive-value to have. Cryptocurrencies are not only a negative-value option, but also waste vast amounts of electricity and computer hardware in order to provide it.
Cash is significantly harder to handle and move, while being easier to track.
And I agree with you there. State action is useful primarily to solve coordination problems and handle negative externalities. The existence of bitcoin markets is a pretty significant negative externality due to the coercion it facilitates.
Having a third participant that you can sue or subpoena is the important part. Is there a good word for this sort of thing? Facilitated by a third party with significant assets subject to the legal system? I heavily favor using such parties for my personal financial affairs - it's not that I trust them more, but it's that I have more recourse when things go wrong.
If you don’t own crypto this can’t happen to you. What other things are you afraid of happening to me? Can you please ban me from visiting the Middle East?
I have the option of exchanging money for bitcoin and then sending it somewhere. Lots of ransomware includes detailed instructions for how to do this, along with phone support to walk you through this process. This isn't a hypothetical, and I don't own crypto, and I don't want someone to encrypt my hard drive and then give me instructions for buying crypto and sending it to them.
You do realize that the act of encrypting your hard drive and demanding ransom for it is illegal, right? So requiring illegal crypto for the ransom is no obstacle.
You're missing the point - it's not about making receiving the cryptocurrency illegal. It's about making buying the cryptocurrency illegal or otherwise infeasible. The unit economics of ransomware stop making sense when more of the victims are unable or unwilling to pay. Making cryptocurrency illegal makes ransomware victims less able and willing to pay the ransom.
So malware writers will switch to different techniques to obtain ransom. For example, there was an online illicit marketplace akin to silkroad that ran years before cryptocurrencies were even invented. It used Western union money transfers, if I'm remembering correctly, and ran for years before law enforcement was able to shut it down.
Western Union transactions required someone be physically present to pick it up, so you could catch them. How do you freeze a Bitcoin wallet or find who is using it in a straightforward manner?
I’ll just have you drop a suitcase of cash in a high foot traffic area with no camera. Much smoother than it sounds, nobody pays attention. You’ll still have to do a few other illegal things in any case though.
At human civilization's scale, the energy production/consumption market is not a zero-sum-game. I would be curious to know how much extra energy production the crypto-currency world has caused/incentivized to be established (e.g. increased capacity factors of underutilized hydro-power, etc.). It's not as if consumption will grow until there is no energy left for anything else. We will just produce more (which can also be a problem - e.g. CO2 and such).
I don’t think anyone is presenting it as a zero sum game. The concern is increased pollution. In a way it would be much better if it were a zero sum game, as that would mean that bitcoin doesn’t increase pollution levels.
What most people that keep repeating the energy consumption meme fail to realize is that the amount of electricity spent to secure the network is highly correlated with the block reward. Currently the reward is 12.5 BTC/10mins or close to $100k and it will keep getting halved every 4 years. Right now the fast rate of adoption drives the price up but in the long run the price will hit a plateau and the energy consumption will start to decline.
No, mining gold is not cheap. And if by "traditional money system" you mean fiat then I must argue that it is not comparable. Comparing bitcoin to fiat is like comparing gold to paper.
He say it takes 2.55 GW now and "potentially 7.67 gigawatts in the future" because "Economic models tell us that Bitcoin's electricity consumption will gravitate toward the latter number"
I'm not sure what 'economic model' makes him able to predict that mining will triple from where it is now.
This would also imply 6.7 Billion dollars per year just to buy the btc that the miners generate and around 6 million mining rigs running at 1300W.
I think the economic model here is that until there is no marginal increase in profitability from increasing mining capacity, miners will continue to increase their capacity.
On May 15, 2017, global hashrate was 3.6 TH/s. Six months later, on November 15, 2017, global hashrate was 10.3 TH/s (source: blockchain.info). On May 14, 2018, another six months later, global hashrate was 30.1 TH/s, an increase of just under 3x in each 6-month interval. While past performance cannot definitively determine future performance, it’s a pretty strong trend.
... than fiat must according to the same line of reasoning also consume resources when it is placed in circulation. That's because if new money $X has value Y, then players in the economy will spend up to Y in resources to be first in line to get $X when it is issued.
One way this may occur is via wasteful government spending (proof of waste?) since the usual path involves the purchase of government debt with newly issued currency and then its entry into circulation via government spending. Another mechanism might be complex financial shenanigans to extract small gains from the cascade of economic transactions (HFT, etc.) that occur as $X is initially lent by banks.
All these activities consume vast amounts of energy. I'd like to see someone calculate the energy consumption of the entire fiat currency system including central banks, the financial industry, and government enforcement for starters.
Furthermore according to the reasoning put forward in the above essay this activity must be "useless," or its usefulness would be priced into it and X more of it would be done where X is the value of its utility. How useful is high frequency trading? How useful is government spending on weapons we will never use? These activities are as useless as Bitcoin mining and may be more harmful in other ways. Warfare kills people.
In other words if the reasoning in the above article is correct then money may be intrinsically expensive. Gold is issued via gold mining, etc.
Tangent: if this is true then this explains the apparent political and economic impossibility of small government under current regimes. We need a monstrous, bloated government to burn the work required to issue money.
So there's not even a logical organizing principle that would actually benefit someone, it's just collective insanity. There's still no obvious application of Bitcoin. This is just slavish alter building to the god of markets.
His findings are not completely accurate, but also not completely wrong either. My own formal well-documented estimate is that Bitcoin used 0.09% of the world's electricity as of January 11th, 2018 (http://blog.zorinaq.com/bitcoin-electricity-consumption/). After 4 months of growth I estimate we are around 0.15-0.20% today. Therefore reaching 0.50% by the end of 2018 is plausible; not likely but plausible. That said, overall I think Alex de Vries's figures are rather inflated because he makes two errors. Firstly, his model assumes that manufacturers sell at cost and make zero profits, which allows miners to spend more on electricity. In reality we've seen Bitmain, the largest manufacturer, make billions of dollars over the last few years. No one expects them to make zero profits by the end of 2018. Secondly, de Vries makes the mistake of assuming that all cryptocurrency mining chips produced by TSMC on behalf of Bitmain end up in Bitcoin mining machines. In reality many chips are fabricated for cryptocurrencies other than Bitcoin. This altcoin sector is rapidly growing. It wouldn't be unreasonable to assume that Bitcoin mining machines will represent less than half of Bitmain's sales by the end of 2018.