700kWh per transaction? That’s absurd amount of power. That’s 70 EUR of energy per one transaction at current (EU) exchange price.
Is there anyone here knowledgeable enough about this issue to say whether those numbers are correct, or just an overestimate? It feels wrong.
They are insanely off, yes.
The number varies a little bit (I’ve seen estimates 600-1200 kWh) but this is well within an order of magnitude of being correct. It’s the nature of the competitive mining network and the proof of work system: if you can spend more computing power (i.e. energy) than everyone else there are lucrative mining rewards to be had. At the same time adding more computing power to the network doesn’t add more transaction processing power, because mining difficulty is constantly adjusted to keep the speed more or less constant.
This naturally leads to exorbitant power consumption per transaction. Note that most of this power is not being purchased at EU exchange prices (mining naturally moves to where electricity can be had for cheap to maximize profits).
I just hope bitcoin will finally die. It’s literally just wasting absurd amount of energy, only to allow scammers to scam billions of dollars from victims, and regular people to steal from eachother by investing into it. I mean, if the only use of bitcoin by now is for speculation and investment, then it means that any dollar you made, you literally stole from someone else who will be left with useless bitcoin once it’s all over. There’s no value, and with the ledger getting bigger and bigger, and bitcoin more expensive to mine, it will eventually be worthless. And we all know it, so anyone who makes thousands of dollars, there’s someone who probably financially ruined himself by making a wrong and stupid investment at the wrong time.
I hate crypto so much :D.
You can exchange it for goods and services in the same way as the dollar. Which is the goal of it in the first place. Disregarding the cost for the sake of this point, it functions in a similar way as the dollar, which you could argue is also just used for speculation, but it would be equally inaccurate. Then there is describing bitcoin as all of the cryptocurrency ecosystem, which is also incorrect. It’s an evolving technology, and the system bitcoin uses is legacy and expensive, and is currently being kept alive by being the first in the space, money interests wanting to keep it dominant because of investment and a horde of cultlike followers. However, in the ethereum ecosystem, transactions keep getting cheaper through layer 2 protocols and upgrades to the system. It uses proof of stake which is vastly cheaper. I think there certainly are valid arguments against cryprocurrency, but the stuff everyone keeps NPC copypasting is generally nonsense.
You can exchange it for goods and services in the same way as the dollar. Which is the goal of it in the first place. Disregarding the cost for the sake of this point
But you can’t disregard the cost. No big box retailer is going to start accepting BTC transactions at the point of sale. It takes too long, and is too resource intensive. So you’re essentially limited to secondary markets or person-to-person transactions where seconds in processing the transaction don’t count. Not to mention the volatility of the exchange rate. A business or a government aren’t going to accept that the currency they accepted today could potentially be worth half its value tomorrow. Without some sort of major catastrophic world event, something like that isn’t going to happen to the dollar. That kind of shit destabilizes entire economies and nations.
There’s also the issue of reputation. Regardless of validity, cryptocurrency is seen as something used for criminal activity. If your local city government said they’d start accepting BTC for your water bill or trash bill, the majority of average folk are going to think “that stuff people buy drugs with on the Internet?”
BTC has never been and will never be viable as a mass market currency system.
You are right I shouldn’t have equaled bitcoin with the rest of the crypto ecosystem. While most crypto is utter scam, it’s true that there have been some slight advances here and there, and there are coins that may be actually useful for some cases, mostly Monero and I suppose Ethereum. I’d still say that crypto has done more harm than good in the world, and I say that as someone who’s really focused at privacy, care about it a lot and have invested significant amount of time and effort into staying as private as possible.
But it’s great that Ethereum managed to solve most of the issues with Bitcoin - unless I’m mistaken, it’s not really used for investment speculation, and if it managed to keep the energy requirements low, that’s good. But last time I remember researching about blockchain (it was few months, so feel free to correct me), isn’t it running into serious issues with ledger size, that makes it infeasible for long-term (decades) of use, without sacrificing some of it’s guarantees? Which is one of the main issues with blockchain tech in general, that I don’t think has been solved so far.
It’s incorrect. I would comparing to fuel consumption in a car based on how many times you turned. If you make more turns on your way, it would seem your car is more efficient, when in reality there’s very little relation between turns and fuel usage, just like there is little relation between number of blocks mined and transactions.
It’s an overestimate. Right now the total cost per transaction is around 82 US dollars, which at current exchange rates is 75 Euros. That cost covers everything - electricity, rent for the building, salary for staff, taxes, depreciation of mining equipment, and whatever profit is required to keep the miners in business. I don’t know what proportion of miner costs actually goes to electricity but I expect it’ll likely be much less than 70 Euros.
Perhaps someone got that 700 kWh figure by doing the reverse calculation - looking at how much a transaction cost and then assuming that all goes toward electricity.
Ethereum transactions are claimed to use about 1-22 Wh per transaction. Not sure where the wide range comes from, but at least it’s quite a lot better then Bitcoin’s ~700kWh per transaction. Ethereum is comparable to how much a credit card transaction is said to spend, although those figures only take into account power needed for their computer systems. Blockchain currencies replace a lot more infrastructure than just the computer systems, so I think it’s reasonable to say that Eth2 is way more energy efficient than credit cards.
That’s not enough to make it a replacement for credit cards yet, but it’s a good lowest of the low barriers to be crossed to qualify as a replacement.
In a few years, we’ll probably be spending a huge amount of power for AI also, and there doesn’t seem to be any Proof-of-Stake -like technology to help with that. Good times.
Lol
Indeed, sir.
Because of these transactions, many countries, such as the United States, could face freshwater shortages if the currency becomes more widely adopted.
False.
Blocks get mined (secured) with the same amount of power no matter the number of transactions in each.
Interesting that an article like this would come out right as Bitcoin’s value is going up and the US SEC is considering approval of several Bitcoin ETFs.
It’s a necessary sacrifice to push more ponzi schemes
It’s actually a scam pushed by Big Water to get you to use more water and line their pockets.
When the alternative to prove of works (vouched by those hoarding compute resources) is prove of stake (vouched by those who can afford to park piles of money), both are suck for their own reasons.
False. Mining is what uses electricity (and water) in bitcoin, not transactions. Adding more transactions does not add to the cost. Calculating consumption per transaction is misleading as the two are not related.
What does add to the cost is complexity, and complexity is calculated based on number of miners in the network in order to achieve the sweet spot of 1 block every 10 mins (if i remember correctly). If there’s a lot of competition, each miner will have to use more electricity to win.
Finally someone that get’s it
These data centers consume water for cooling systems
How does a data center consume water? Doesn’t every liter that enters as freshwater leave as slightly warmer freshwater? What am I missing here?
Some of the water is evaporated so it doesn’t leave as a liquid.
That’s unlikely in a closed heat exchange system. Maybe some additional evaporation because the water is slightly warmer. But unless I’m missing something, it seems very misleading to suggest that a Bitcoin transaction uses 16 kilolitres because of evaporation. Napkin math, it would require about 10 megawatt/hours of energy to evaporate that much water (please correct me if I’m wrong). I’m not a Bitcoin fanboy, I just don’t like BS.
someone from a totally different thread mentioned that the water can’t stay in the system because of whatever mineral stuff from the cooling pipe/anti-algae/anti-corrosive has to leave the system after certain cycles. So unless you have a treatment plant down stream it’s not exactly “drinkable” freshwater. (and I doubt water regulation would allow that to happen.)
The consume here means that water is not usable for other application. How? I don’t know, maybe it can be used for power wash?
It probably is still a lot easier to make potable than sewer water or even river water though. At lease you know exactly what contamination is in it.
Water used to cool data centers is either consumed, meaning it evaporates into the atmosphere via the data center’s cooling towers or discharged, as industrial wastewater, usually to a local wastewater treatment plant.
It can’t just be dumped into a river, has to go to a sewer treatment plant.
edit: They do recirculate it, but it eventually needs to be replaced. And some facilities have treatment plants on site, so doesn’t necessarily needed to go to a sewer treatment plant.
Some water is used in humidifiers, there are also systems that use direct evaporative cooling where the water is eveporated to cool the hot air. There are probably other ways the water is lost.
AWS’ preferred cooling strategy for its data centers is known as direct evaporative cooling. In this system, hot air is pulled from outside and pushed through water-soaked cooling pads. The water evaporates, reducing the air’s temperature, and the cool air is then sent into the server rooms.
Even if so, does evaporating cause it to exit the water cycle?
These cooling systems remove and release all of the heat produced inside a data center – from servers, IT equipment, and mechanical infrastructure – into the outside environment, through a cooling tower that uses a water evaporation process.
It goes outside and eventually becomes rain.
Aquafers do not refill as quickly as industry sucks them dry. It’s not just a Bitcoin or even a cloud computing problem, but the author is using this fact to make Bitcoin look even more ridiculous.
Swamp coolers.
Fans blow over water to lower the pressure, causing evaporation to occur at room temperature.
Evaporating water absorbs heat from its surroundings without raising the water’s temperature as it undergoes a phase change. It absorbs nearly 20 times more heat than it would from being heated from 50 degrees F to 100 degrees.
i strongly urge skepticism when reading articles about the environmental impacts of bitcoin. I am not saying that bitcoin is a sensible use of resources - rather that the claims made about the environmental impacts are often overstated and based on models extrapolated to absurdity. For example, see https://doi.org/10.1038/s41558-018-0321-8 where Mora, Camilo et al suggested that “Bitcoin Emissions Alone Could Push Global Warming Above 2°C”. Then read Implausible projections overestimate near-term Bitcoin CO2 emissions by Masanet et al.
Again - the environmental impacts of cloud computing in general and bitcoin in particular are something we should be concerned about. But there are a number of researchers who have made wild claims that should be treated with a critical eye.
Yeah, generally miners will set up in places with cheap electricity. And excluding places like Azerbaijan, those sources are generally renewables.
Yeah. Right now, the cost of a Bitcoin transaction is around $65 US. That price includes all of the expenditures that the miners have made on resources (electricity, water, rental costs for the space they’re using, hardware depreciation, etc.), as well as whatever bit of profits it takes to keep miners in business. That puts a cap on whatever environmental impact the transaction is having.
So if it costs $65 for a transaction then why isn’t that the transaction fee? people would be loaing money if cost is more than the fees
Bitcoin is inflationary, it’s generating new Bitcoin with every block and issuing that to the miners. That new Bitcoin combines with the transaction fees to pay the miners.
So mining is the bulk cost not the transactions. because my last bitcoin fee was like $10 or something
Yes. If you’re wanting to know how many resources mining a transaction takes, that’s the value you need to look at. The block reward effectively goes into subsidizing the transaction fees that are being paid.
Bitcoin is deflationary. There is a hard limit on the total number of bitcoins that will ever exist. Every so often, the reward for mining a block is halved. Eventually there will be effectively zero reward for mining at all.
Maybe in the long run. However, when you want to actually calculate how much each transaction costs, you need to account for the fact that right now Bitcoin is inflationary. It won’t stop issuing new tokens until around 2140 AD, assuming no hard forks happen to modify that issuance strategy in the meantime.
(Checking if outbound federation is back)
Yeah, if they had said 10 gallons, I’d buy that, but a whole swimming pool of water would be worth far more than a transaction fee I’d expect.
I don’t understand some things in the water consumption.
Why do they need to humidify the air for the datacenter?
Why is there water consumption for cooling? Aren’t they recirculating water used for watercooling? Or are they using f*ing tap water then throwing it out?
Water for electricity production, kinda, yes. Could be indirectly attributed to their water consumption as they are using the electricity produced by the sources using water.
Why do they need to humidify the air for the datacenter?
Static electricity. Humidified air dissipates static charges before they can build up enough to arc and cause damage to sensitive components.
I worked in a “datacenter” where the humidifier function for the HVAC unit was turned off because it leaked under the floor into an adjoining office when it was trying to humidify. Management refused to fix the unit due to the cost, and saw no issue with running the room with relative humidity in the teens all winter. Madness.
If we didnt run a dehumidifier the humidity in my house would stay above 80% most of the year. We have a decently large dehumidifier and by itself it cant get it below 45%. But 45% is much better then 80+% so it could be worse
Yes of course but you have humans in the house who exhale H2O all the time. A datacenter doesn’t have many of those (per square metre or foot or whatever you measure your datacenters in)
The vast majority of it definitely comes from the area in which I live.
Temperature is the more important factor. Even if the ambient air is at 100% humidity, if it is very cold, the relative humidity after heating it will be very low.
Yeah and too much is bad too, because it can condense on coolers. I’ve been told it’s also bad for bearings in HDDs when it’s too low.
Of course in a datacenter being low on humidity is a much more likely thing as there is nothing to emit humidity (e.g. humans).
Not only is the science underlying all these findings completely non-existent, they only “guesstimate” what the water usage of what every thing that uses water is; then blindly divide that by the transaction volume per time period.
Not only is that method highly flawed; it’s incorrect. Computers do more than mine crypto; and 1 transaction typically costs not even 1 tenth of a percent of most miners’ overall computer resources. This is due to the fact that many miners are utilizing either a GPU or FPGA style device to power optimize and optimize the mathematics necessary to secure a transaction.
That might have been true a decade ago. But GPUs and FPGAs have long been obsolete for mining Bitcoin.
Mining is happening on custom silicon in large-scale operations. They specifically observed several of those large-scale operations in multiple nations and extrapolated out. I don’t see how that methodology is flawed.
Bitcoin is fake so fake story of fakeness.
These calculations are a bit off IMO. They factor the total amount of mining and divide it by the number of transactions.
However, the amount of mining is not dependent on the amount of transactions.
I’m not a fan of bitcoin due to the wasteful proof of work mechanism but ‘blaming’ the transactions is not really fair IMO, especially because people don’t really use bitcoin as a payment method anymore. It’s just used by speculators now.
Wasteful and unsustainable.
However, the amount of mining is not dependent on the amount of transactions.
Entertain my ignorance on this for a second, but isn’t there some sort of dependence here? Like not a strictly casual dependence, but if transactions were, say, to magically halve for a few days, would that not affect the mining required and thus the total energy expenditure of the mining?
(Obviously the limit case would show this to be true, in that in the absence of any transactions at all, mining would cease. But I’m after something a bit more clearly casually related, somewhat like supply and demand in the marketplace – consumption of beef driving more supply and more methane, e.g.)
The mining isn’t based on transactions, the mining is set to target a 10 minute block time, the difficulty changes based on how much mining capacity there is.
If a block takes less than 10 minutes the difficulty goes up. If it takes more the difficulty goes down.
Each block has a maximum capacity of 2500 transactions, which is why gas is used to prioritize who gets added in a block.
There have been many blocks with zero transactions.
(May bot be correct #)
No, transactions piggyback on the mined blocks but if there are no transactions mining still happens.
I see. But in the limit case where just everybody decided BTC is nonsense and stopped transacting entirely, while mining could continue, eventually it would die out, right?
So in a sense, do transactions not drive the need for mining? If that’s the case, the connection isn’t directly casual so much as one of complicity. Does that make sense or am I still barking up the wrong tree with this way of thinking?
I see. But in the limit case where just everybody decided BTC is nonsense and stopped transacting entirely, while mining could continue, eventually it would die out, right?
Probably, yes. But it’s important to realise that bitcoin started as a payment system. Meaning lots of daily transactions would be done. These days it’s used more for speculation, as a value storage system and for transferring to other coins. Which implies a lot less transactions.
Everybody basically has already decided that BTC is nonsense for payments and stopped using it for that. This is exactly why the transactions have so much “overhead” because so few transactions are compared to so much mining.
So in a sense, do transactions not drive the need for mining? If that’s the case, the connection isn’t directly casual so much as one of complicity. Does that make sense or am I still barking up the wrong tree with this way of thinking?
Not really, no. Miners mine as an investment. The whole payment system community has already been taken over by other systems which are much better suited for that, like Etherium, which has proof of stake for low overhead, fast transaction time and smart contracts. Or Monero, which hides the identity better so it’s more common in the purchasing of certain illicit substances.
The BTC community reacted with lightning but it was too little too late to solve this usecase. The BCH (bitcoin cash) fork was also motivated by this as far as I understand, because the miners were opposed to any changes to facilitate easier payments and lower transaction costs. But this is more hearsay I have to admit, I’m not a cryptobro and not fully into this.
So now BTC is less like a “bank” and more like a “goldmine”. That’s how you should see it. Even though gold is a useful material, most investors that buy gold don’t buy it with the intention to ever sell it to a factory making connectors or whatever. They just buy it because it’s scarce and the price keeps going up, and people assign value to it.
Bitcoin is in the same position now. It has value because people decide it has value. This is not really related to its potential use as a payment system. Miners (the ones who control the bitcoin stack now) are not even interested in its use for that purpose and seem to actively block enhancements to make that easier. Though transactions are still necessary for trading between investors, it’s much much less in volume than it would be if people were still using it to pay for stuff in shops.
But anyway, going back to my original point: Articles like the link here that claim that Bitcoin is a really shitty payment system are kinda stating the obvious. It’s practically speaking not a payment system anymore even though it technically could be used as such.
Not even a mention of lightning? I have no idea if it works as I’ve been hearing both yes and no for several years, but writing such an article without mentioning what at least theoretically would be the solution just seems bad.
Last I saw Lightning was pretty much DOA, it’s been around for many years and almost nobody’s using it. At the time I was checking there was an order of magnitude more activity transferring Bitcoin on Ethereum using WBTC tokens than using Lightning on Bitcoin itself.
Yeah I made a bitcoin payment recently and I was very suprised to learn that my wallet (CakeWallet) doesn’t support Lightning payments at all. So I had to do it the old way.
Very weird because bitcoin advocates always pushed it as the holy grail. But I guess bitcoin as an actual payment method is just really too niche for it to take off.
I recall back when Lightning was first come up with thinking how incredibly hacky a solution it was, full of awkward workarounds for the limitations of Bitcoin’s blockchain. A few small changes to the blockchain would have made it so much simpler and more robust, but at that point Bitcoin’s immutability had become such a fundamentalist religion that any such changes were absolutely rejected. They wouldn’t even fiddle with the block size, let alone consider expanding its scripting capabilities.
Then Ethereum came along with the exact opposite philosophy, it’s willing to continue making changes to the foundation layer with the overall goal of making Ethereum more functional for diverse applications. Ever since then it’s just been a slow transition of everything useful moving over to Ethereum and Bitcoin becoming ever more insular and obsolete by comparison.
It seems like I haven’t thought about Bitcoin in years. When this article came up it took me a moment to shift the mental gears and go “oh yeah, that. I guess it’s still around.”
As tradition I won’t read the actual article and only comment on the headline - while BTC is a massive energy waste, it seems unlikely that each transaction would waste so much cooling water. Maybe each mined block, but each block should contain thousands of transactions
OP, your “news” article sucks. It’s just clickbait
