Bitcoin's thirst for water is just as troubling as its energy appetite It probably won't come as a shock to those aware of the electricity usage of the Bitcoin network, but the world's most popular cryptocurrency also uses tons of water to keep itself running. According to Digiconomist founder and Vrije University data science PhD candidate Alex de Vries, the Bitcoin network may be using as much …
I see both as the same.
T. Pratchett said it well in "Making Money" that even if it takes more money to mint a coin, that coin goes on doing its job for a long time. A Bitcoin, on the other hand, has to be remade every single time at a far greater cost than the value it represents. That cost includes power plant emissions (directly and indirectly) and fresh water. There are whole giant mining farms that have been abandoned when the price of crypto has dropped too low. All of the bespoke equipment has loads of energy embodied from what it took to manufacture, now nothing but scrap since it tends not to age well.
You're totally misconstruing the operation.
First, a (fraction of a) bitcoin is the reward given a miner for appending to the bitcoin blockchain. That work of appending secures recent transactions on the chain, as well as attesting to new ones. Certainly, for a person to transfer--"use"--the reward, additional work is required, but that's actually as true as for any form of currency.
Now, as for the claim that compute farms have been abandoned, I find that REALLY hard to believe. This isn't 2000. The market for compute has tremendously expanded, and it seems incredible to suggest that someone owning an unexpectedly idle bit barn would not figure out a way to make use of that hardware.
"it seems incredible to suggest that someone owning an unexpectedly idle bit barn would not figure out a way to make use of that hardware."
For what? It's very dedicated hardware for which there isn't any other use.
There have been plenty of stories about abandoned bit barns. The price of a crypto drops too low and the mining company goes BK and/or can't pay their power bill. By the time a court might get around to ordering the disposal of assets, the gear has been supplanted with something better. Not there would be a market for the old stuff if mining doesn't make enough to break even.
Here's a story on Forbes: https://www.forbes.com/sites/colinharper/2022/09/30/bitcoin-minings-first-major-bankruptcy-creates-uncertainty-for-key-partners-opportunity-for-others/?sh=662322ec657d
The electricity expended to move a few bitcoins in 10 minutes is enough to move a ton of money around in the credit card system near instantly.
Physical currencies can be spent with only the energy it takes to keep two humans awake for a half-second, if they're already close enough.
You've not thought this through.
" A Bitcoin, on the other hand, has to be remade every single time at a far greater cost than the value it represents. "
No, a Bitcoin, once made, lasts forever. It can be 'lost' (eg by someone losing their crypto keys), but it cannot be 'destroyed'. And the cost to make it is, by definition, less than it's value (otherwise no-one would mine). It can be transferred infinitely many times, although of course every transfer incurs a cost. None of this is much different from fiat currency except the effort it takes to mine the bitcoin in the first place (and this effort DOES consume a LOT of electricity but not water - that is a bullshit claim as water *used* by data centers and power plants for cooling is not *consumed*, it is released again as slightly warmer water).
" All of the bespoke equipment has loads of energy embodied from what it took to manufacture, now nothing but scrap since it tends not to age well."
That's not much different to a lot of IT hardware. The hardware is mostly similair to graphics cards so can easily be reconfigured for, for example, animation studios. Although it is certainly less long-life than a general-purpose server, it's probably about the same longevity as the average mobile phone. Not that it's in any way good, just that there are surely many possible ways of repurposing mining rigs to, say, weather forecasting or AI workloads, and the reason it's not done is the same as for most other modern technology - it's cheaper to buy something new than to repurpose something a couple of years old. But that's got all to do with modern technology design, manufacture and supply chains, not specifically to Bitcoin mining.
Yes. The miners had to move from GPUs to ASICs years ago to stay competitive as the search space grew. As mining becomes unprofitable on anything other than the latest gear, previous generations are just so much fused sand.
If there were lots of old GPUs that someone wanted in abandoned mining rigs, they'd be pushing GPU prices down. That's not happening; therefore abandoned mining equipment isn't useful.
I think the only time the used GPUs pushed down GPU prices was shortly after China made it illegal to mine crypto, and even then if wasn't by much. In fact, the crypto miners tend to, if anything, increase the cost of more general-purpose compute-graphics GPUs by purchasing the latest ones every chance they can get.
But crypto innately cannot be mass adopted, the system is deliberately inefficient. If the entire world ran on crypto then the energy and water demands would be impossible to meet.
Not to mention all the other issues, like achingly-slow transaction times and changeable gas fees.
Ah, the "adoption" handwave. A great way to demonstrate that you're not willing to consider the actual practicalities.
Why can't everyone have enough food and water? There's plenty to go around. It's just a question of distribution.
(If only we could find a way to harness the power of sophomoric argumentation...)
This post has been deleted by its author
This post has been deleted by its author
Again, the water is not destroyed. Whatever evaporates forms clouds (real ones), and then rain somewhere downwind where it most certainly _can_ be drunk.
Yes, if there are local use issues, then _any_ data center should be managed like any other large user.
Are you going to complain about the water used by sewage plants? Its an almost identical calculation.
It's a combination of the water consumption for the generation of the power consumed and the water consumed per kWh for cooling the data centre. They don't actually know the specifics for the latter for mining facilities but treated them like any other data centre, adjusting for geographic variations in water consumptions (e.g. making use of evaporative cooling where the climate permits) based on the location distribution.
As far as the paper goes, there doesn't appear to be anything specific to cryptocurrency mining that would result in higher water consumption than any other data centre operation of similar power consumption. That said, whether this is because there really is no difference, because the authors lacked the capacity to investigate fully or because they didn't consider this question is unclear.
Your second paragraph is irrelevant to the study. The study considers water consumption for Bitcoin mining, not how that compares to water consumption by other types of data centers. There's certainly the implication in the discussion around the study (I haven't looked closely at the study itself) that this question is of interest because Bitcoin mining returns less utility than other data centers might; but that's peripheral to the study itself.
I think you might misunderstand. I wasn't suggesting that the comparison (as in, presenting a rundown of the difference between applications) between other types of data centre is needed but, rather, that the data is potentially inaccurate because it is derived from general information on data centre water consumption, with the assumption that there is no difference. This seems like quite a sizeable assumption to make and, while the data is probably somewhere in the right ballpark, it could still be off by quite a bit.
That was my big question as well.
Bitcoin mining does not directly consume water, it can only do so as a by product, meaning energy generation or datacenter operation.
Moreover there is a gigantic falsehood here - none of these operations *consume* any water!!!
If a farm is using water, that water is consumed - It is taken into the structure of plants or drunk by animals, and is no longer available for other purposes (or at least, not for quite a long time as it cycles through the food chain / waste and back to water)
If industry is using water for making stuff (including food industry for bottling or making drinks), that is also consumed.
Water that is used for cooling is taken as an input cold, and then released at a higher temperature. The water is *used* but not *consumed* - it's still available to anyone downstream. Similairly if it's being used for hydroelectric power. In other words the whole "Bitcoin consumes XXX water" is a giant bollocks
If you pump water out of a river then turn it to steam, it's unlikely to return to that river in any rapid amount of time. Particularly if you're actually pumping an aquifer instead. It's an issue Phoenix Arizona is struggling with due to its arid climate.
Returning heated water to a river can sterilise it if the river temperature is already high, by causing oxygen to escape out of the solution.
Moreover, while plants/animals do consume water, they use it as a solvent for internal chemical reactions within cells and subsequently to expel waste. That's a good portion what goes down the u-bend when all is said and done.
Yes, you've missed something somewhere.
Water as a utility output — filtered, treated, and transported to point of use — is a commodity and is far from the thermodynamic equilibrium for water on Earth in general. Water that has been put to industrial use, such as cooling a data center or being similarly put to use in generating the power used by that data center, will inevitably be less-ordered and closer to equilibrium. It will be warmer; some of it will have been lost to the atmosphere and probably some to ground as well; usually it will be contaminated to some extent. Entropy has increased.
Providing more water to those points of use for continued operation thus requires energy and raw material (dirty water) inputs.
Therefore while there may still be the same number of H2O molecules leaving the system as entered it, the actual point-of-use resource — treated and transported water — has been consumed.
I'm not sure why this patently obvious process is obscure to so many commentators here. There's a reason why businesses (and individuals) pay water utilities: they supply a commodity which is not freely available. Putting that commodity to use consumes it. Some may be recovered and recycled back into the input stream for downstream consumers, but that can be factored back into the cost of consumption unless the consumer is well outside the norm.
This post has been deleted by its author
Insiders? There is nobody *inside* Bitcoin. It is not a corporation or business. It's a decentralised bundle of specialised maths and computer science that is constantly balancing itself. Holding Bitcoin, if you boil it down, is hedging against human derived financial constructs (which are just as artificial, if not more artificial because changes to it are arbitrary) and betting on cold, hard, fractal maths.
If I was forced to make a choice between putting my life in the hands of maths or a bunch of humans I've never met that don't care about me, I'm going with maths...because at least with maths I have a shot at understanding it and it's a level playing field that nobody can influence...maths doesn't suddenly change because someone somewhere has decided to change the goalposts. If someone decides to fuck with the maths, then that new maths becomes it's own chain. The original chain is still there and ongoing while the "new and improved maths" that is usually created by a person to create some sort benefit for themselves, ends up on it's own chain...and if that maths sucks, then the chain withers and dies and the value from that chain ends up back on the original chain anyway...because maths.
Bitcoin is basically a drawn out war between people and maths...as far as I am aware, throughout the history of mankind, humans have never beaten maths.
No, it's go nothing to do with maths, BitCoin's value is the perception of the worth of a finite resource, or to use a word you may understand; greed. It similar principle to the value given to gold but that also has real uses, so far more like 17th century tulip bulb bubble and every other commodity fad that comes crashing down when a small percentage of the traders finally see sense.
Is the inference that certain computing activities should be banned because they are unnecessary? That would include all gaming, streaming and much surfing. But nothing that politicians do, of course. Everything politicians do is OK. Like all that flying across borders during Covid, Macron visiting the Olympics in Tokyo etc. Just the stuff the proles do needs to be banned.
Activist science is usually bad science, and the use to which it is put should be questioned.
Is the inference
Is it? You're the reader; you create the inference.
that certain computing activities should be banned because they are unnecessary?
I don't believe anything in the article implied that, no. The study's author did suggest that the resource consumption of Bitcoin mining could be greatly reduced, and that mining did not offer utility.
I'm ignoring the rest of your irrelevant rant.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
