Bitcoin Forum

.

Thank you for this very interesting calculation. It's a nice addon to a thread I started quite a while ago that led to a fruitful discussion. See https://bitcointalk.org/index.php?topic=181759.msg1897702#msg1897702 (https://bitcointalk.org/index.php?topic=181759.msg1897702#msg1897702)

I use your calculations to figure the upper boundary on electricity usage. However, I suspect the lower boundary on cost per KWH is about $0.03.

Over 50% are paying less than $.10 per KWH: https://bitcointalk.org/index.php?topic=333720.0 (https://bitcointalk.org/index.php?topic=333720.0)

I remember some Indonesian miner guy that was paying $0.03 per KWH.

Cost in Iceland per KWH according to google is $0.043.

I can buy solar panels at $1000 per KW. Assuming a 5 year effective life span at 10 hours of direct sunshine per day = 365 * 5 * 10 = 18,250 KWH. Divide that into $1000 ($1000/18250) and we get $0.054 per KWH.

Personally, I would just use $0.05 per KWH. At around $600 per BTC and assuming 3600 BTC per day, we get:

$600*3600/$.05/KWHR= 43.2*10^6 KWHR for an <b>UPPER BOUND</b> on power usage.

Remember, as BTC usage goes up world wide, power usage will stay about the same. I suspect the cost of using Fiat (especially when factoring in both the hidden and apparent cost of inflation) is much much higher.

Very nice "rule of thumb"!  As an "absolute lower limit" you could use $0.015 / kWh:

Price of electricity = 0.015 USD/kWh = 15 USD/MWh

So we have 3600*X = P*24*15 or P/10=X

first, you can only buy panels by peak power [kW], not by energy [kWh].

second: would you really invest on expensive mining gear and then run it only when the sun shines? Your gear is outdated quickly, so amortisation time is short.

This is just wrong. Power usage rises with the valuation of BTC:

GlobalEnergyConsumptionForMining ~= (MiningReward + TransactionFees) * bitcoinValueInUSD / EnergyCostInUSDperkWh

most certainly wrong.

please read my thread carefully where you will find your arguments proven wrong (see signature)

I fixed my typo

True, but power in Iceland is still $0.043 per KWHR

Value of BTC may go up, but mining reward is definitely going down every 4 years.

What is the cost of using fiat? I see inflation as a hidden tax equivalent to the amount of money being printed. I lean toward Austrian economics though. If you have a different economic philosophy, we will probably never agree on the cost of using fiat. But here is my take on it: If the USA prints 10^12 dollars per year, then this would be equivalent to 10^13 KWHR (at $0.10/KWHR) per year. This is certainly more than: from your OP statement thread in your signature.

I read part of that thread and my assuptions could actually make things worse. Mining would be done in the areas of the world where energy is the cheapest. At $0.05 per KWHR, then your 7% would become closer to 14%. But that would decrease over the years as the mining reward keeps halving. But even at 14% of world wide energy usage, it seems to me that the cost of Bitcoin would still be less than the cost of using fiat currency.

BTW, I don't think that Bitcoin (or any other cryptocurrency) will ever completely replace fiat currency. I spend a lot of time in Kenya with my youngest wife who uses M-Pesa. M-Pesa is everywhere, you can't walk a block on a business street without seeing a couple signs. She uses it to send money to her parents via text on her phone. She uses it to pay her electric bill. We were in a matatu going past the electric company main office and she was making fun of all the old people (younger than me...but beside the point  :D ) queing up to pay their electric bill instead of using M-Pesa. And everyone in Kenya (even if they sleep in a cardboard box) has at least one cell phone. One day, she wanted to stop for a coke and I mentioned that I hadn't brought my wallet and she had no money. No problem she said, stopped at the next M-Pesa sign and converted some of her M-Pesa on her phone to 500 shillings. The point being is that Kenyans are already using an ecurrency for all sorts of things, but they don't use it for small transactions. Fiat currency is simply faster and easier.

What I do think that Bitcoin will do for fiat currency is force governments to stop inflating it so much. There will be a balance between slightly deflationary Bitcoin and slightly inflationary fiat currency. When a government starts inflating fiat to much, people will just use less of it (and at a much higher velocity) and switch to using cryptocurrency more. This will accelerate inflation and force the government to print less fiat(or find themselves unable to operate effectively).

Just curious.  How many do you have?  One is more than enough for me.

Just Two.

An old cougar that caught me when I was young and naive. As I got older and my SMV went up, I negotiated the right to take a second wife. All three of us lived together in Nairobi for a short time, that was a total disaster. Feminism vs. Femininity...some day I might write an article on it.

Still you can create more green solution. But consumption is still not high enough to worry about

If it was a typo, then the number is wrong. It's not done by buying a PV module. I guess Germany is a good place to look at for PV. http://www.photovoltaik-guide.de/pv-preisindex (http://www.photovoltaik-guide.de/pv-preisindex) says they're at 2$ per Wp.
And it's still not done because you need some kind of power management. If you run in island mode (off-grid, to avoid paying grid fees) you need a battery to get most out of your gear, add another >2$ per Wp.

Or - of course - you let the german public pay: Feed your PV into the grid, get a subsidized price, and then buy coal power to run your rig ;-)

OK. But iceland cannot supply the 7% (or 14%?) of global electricity.

Well, you're right, that's not what I was talking about. I only estimated the energy used for the financial sector.
But if your point is to equal fiat with spent energy, how would bitcoin make things different?


The decrease in mining rewards will followed by increasing gains by fees. Or bitcoin security will weaken.

We can only agree to disagree on this point. Especially, because there are alternatives to PoW.

The OP is nonsense. Here is why.

The amount of energy consumed will be much less than the current banking system consumes which, hopefully, bitcoin will replace. :) This doesn't even take into account the millions of lives (and productivity) saved when nation states can no longer print money at will to finance their endless wars.

If price doesn't increase consumers break down the doors of the electric company and City Hall. Price will increase damn fast when people don't have electricity. I am here in Davao, Mindanao and we are having rotating brownouts. The people and businesses are ready to strangle the necks of the politicians and so yes the price has risen and the new coal powerplants are being built (200MW to come online next year and more coming).

I agree price may not rise if the damn socialists who want subsidies are in control.


Man-made climate and global environmental change is the most irrational thought a person could have. Start here (https://bitcointalk.org/index.php?topic=374873.msg6061621#msg6061621) on the definition of the scientific method.


I assume you are referring to this point:


You must have flunked Economics 101 because you forgot opportunity cost.

They can sell the hardware. Factor that in and you see my point was relevant.

Obviously for ASICs you are correct, but general purpose computers you would not be. That is another reason a cpu-only coin is needed.

Death rate due to coal burning is mostly cancer and respiratory illness.  Mercury toxicity from coal causes neural damage, especially damaging in early development.  A survivable and humane environment is a green environment, AGW notwithstanding.  When economic development is low, you've got bigger fish to fry, but as soon as you can afford intelligent cancer-free children, you're likely to want to spend your developed income on that.

Economic incentives aside, there's really no reason why miners can't be restricted to renewable energy as available.  The network is global, and the Earth turns.  There would always be mining occurring somewhere.  It's only the compulsion to amortize the plant, and to maximize earnings, that compel miners to operate 24x7.

EDIT:  But there is a novel technique for burning coal with iron oxide that essentially eliminates fly ash, if I recall correctly.  It was a research press release, so it might be overblown, and I don't recall much detail -- it was about 2 years ago.

The assumption that the rational actor mines whether or not they can make today's profit and only today's profit via electricity is somewhat flawed.  That would be true in a certain controlled environment and to some extent many miners do wake up and make this decision each day, but.....

In the long term a rational miner makes plans over the course of months including the price of the miner and the price of electricity in relationship to the price of bitcoin.  Excluding the price of miners from your equation would only be true if no new miners came into play ever.  That simply isn't the case.  New miners have constantly and are constantly coming into play.

A person can't just make a theory saying that electricity is the only cost that factors into the price of bitcoin.  That simply isn't the case.  Big companies take into consideration electricity prices, but small home miners are locked into their electricity supply so it is irrelevant to them.  But a large amount of miners are driven into and out of mining much more so by the price of miners, than the price of electricity.  Clever and experienced miners carefully time what miners they buy and when to maximize their advantage.  


---

otherwise I appreciate your insight and thinking this out.  it is a very cool idea

+1

Add 50% to that price/kWh assumption:

http://shrinkthatfootprint.com/wp-content/uploads/2013/09/electricprices.gif

Shouldn't this be then, our long awaited intrinsic value of BTC ?

Also, is everything correct with units ? It looks like they're not cancelling right.

Your equation is (assuming rational miners / no free electricity / continuous network hash) :

N * X = P * Q

N - number of mined bitcoins per day having unit of [ BTC / day ]

X - exchange rate having unit [ $ / BTC ]

P - total network electricity consumption measured in [ MWday ]

Q - price of electricity measured in [ $ / MWday ]

So:

N = block_reward * number_of_blocks_per_day ~ 25 * 24 * 6 (because roughly 10min per block) = 3600   BTC / day

Q ~ 150 $ / MWh  = 3600   $ / MWday

So:

3600  [BTC / day]    *    X  [$ / BTC]  = P [ MWday ]  *    3600 [ $ / MWday ]

Right side units cancel to $, but left side units don't cancel to $. but to $/day ?! Is something missing ?

A future generation of ASICs could be made with reversible gates which means that electricity use becomes far smaller as heat dissipation becomes negligible. (Note that the application suggested here is classical CMOS, not quantum computing).


http://web.cecs.pdx.edu/~mperkows/CLASS_VHDL_99/tran888/lecture003-reversible-logic.pdf
"Power dissipation of reversible circuit, under ideal physical circumstances, is zero."

http://www.ermt.net/docs/papers/Volume_2/issue_5_May2013/V2N5-129.pdf

As the Bitcoin fx rate climbs the incentive to push the envelope in chip design is ever stronger. A SHA-256 hashing circuit is far simpler than a general purpose CPU. Hopefully, Bitcoin mining will be at the forefront of reversible gate technology in a few years making the whole mining network orders of magnitude more energy efficient.

(Mentioned this elsewhere, but had no feedback on how the tech is progressing...)

Price of electric is different around the world. Not to mention all kind of ways to optimize it.

Low power devices will be run with sustainable electricity sources, such as solar cells, making the whole concern about energy consumption an academic debate.

As a dane, let me just say that I am insanely jealous of the cheap electricity everyone else has. It's at $0.45 here.

Price.
Obviously today is as it is now, with a lot of scope for improvement. Looking some years down the road my point is that RLG tech is incredibly efficient because it recycles electricity within the circuit needing only a negligible top-up. While fossil and nuclear prices continue to rise there will be a crossing point where renewable energy sources will out-compete conventional sources.  I think for ASICs this is <10 years away.

Out of curiosity, how much energy does your typical ASIC need? I'm planning to plop down a few windmills on the boat I'm buying soon, and it might be fun to have one running entirely on wind. Just as a proof of concept and bragging rights if nothing else.

Cooling is important factor. You do not only need to produce something. You also need to make sure machine won't overheat.

Interesting part is, even though we usually use electricity for it, there are other solutions.

The "startup" costs of these energy types are very high. As a result one would need to invest a lot of money in order to build a coal or a nuclear plant. It also takes many years to build these types of plants verses a week (possibly) to order, receive and setup a solar panel

Is it really that unlikely that a serious mining operation would run their farm from wind or hydro?

Considering what the mining devices and the infrastructure for it costs, it is not completely impossible in my opinion.

Total cost today (farm deployed):

~$820/TH = $820,000/PH
~$1370/kW = $1,370,000/MW

Annual electricity at $0.88 million at $0.10/kWh _without_ conversion cost, cooling, etc.


At $1.3 million for 1MW at the lower end, it would take less than 2 years to  pay back.

http://www.windustry.org/resources/how-much-do-wind-turbines-cost

If a mining operation takes business serious (= not cowboy style get rich quick mining), it could probably use local resource and government incentives to drive down costs. And after all, if the mining ever stops, the society benefits from all the new wind farms.

(admittedly, I don't know anything about setting up a wind turbine)

Ok, so now you got me curious.

Here's a wind map for the US:
http://hint.fm/wind/

http://www.bneenergy.com/bne-energy-faqs.html

Wind farms are built in places with strong, steady winds to ensure that they produce as much power as possible, as much of the time as possible.  However, wind is variable depending on the time of day and time of the year.  A typical wind farm would produce electricity in the range of 25-40% of its rated capacity over the course of the year, but can produce electricity at full capacity on any given day when the wind exceeds a certain threshold amount.

So the obvious answer, build your wind turbine AND your bitcoin mine where it's windy ;)

Profit from both, Bitcoins and electricity sales.

antirack:  You forgot to answer this question:


This same question applies to solar pannels.

This same question applies to a combination of wind turbine and solar pannels.

I did not forget. My answer was:


So 50% on ASICS and 50% on wind turbine if both would be equal priced ($1,37MM/MW for the mine vs. 1.3MM/MW for the wind turbine).

The big question is where to find the highest mountain with the strongest wind and how to overclock the wind turbine.

 ;D

No, you are wrong, you did not run the numbers.  To maximize your overall profit (assuming you can get immediate delivery of the latest miners) is to put 100% of your money into miners and forget the whole wind or solar energy production idea.  Any other mix of investing in miners/wind/solar will reduce your overall profit.

This is because wind and solar are the most expensive ways to make electricity whereas burning dirt (coal) is the cheapest.

Is that a theory or a fact?

The theoretical 1MW mine above costs $0.88 million at $0.10/kWh per year on electricity cost.

The theoretical wind turbine costs $1.3 million.

It produces theoretical free power after 1.48 years, when it paid itself back.

After that my theoretical mine mines for free.

For a while.

For every single hardware generation I deploy.

Until it breaks down.

If I would be in my home country, where electricity costs close to $0.30/kWh, we have very high mountains AND a lot of wind, how long would it take to pay the wind turbine back and mine for free? A few months.

There is something fishy about these numbers? Double or triple them, it still takes less than a year.

I think that wind turbine will need to receive some kind of maintenance and upkeep every so often.

You also forgot to include the fact that you need to put the turbine on land that you own/control. You would (in most countries) need to pay property taxes on both the land and the turbine. You would also have a lost opportunity cost as if you did not have the turbine on your property then you could have rented your land to someone and received some kind of rent.

You might be right about that.

Found some other info in the meantime:

http://www.forbes.com/sites/christopherhelman/2012/12/21/why-its-the-end-of-the-line-for-wind-power/

According to the federal Energy Information Administration, the “levelized cost” of new wind power (including capital and operating costs) is 8.2 cents per kWh. Advanced clean-coal plants cost about 11 cents per kWh, the same as nuclear. But advanced natural gas-burning plants come in at just 6.3 cents per kWh.

And that is probably for large scale corporate installations by the energy producers, not some single miner with his own land. No word about hydro though.

Back to square one.

Hydro power available at places like Washington State is $0.02/kWh to $0.03/kWh.

I know that the wholesale price of electricity for Xcel Energy in Colorado is less than $0.03/kWh because at the end of each year they have to pay me wholesale price for my yearly solar overproduction (I get dollar for dollar credit for my yearly consumption that is covered by my production).  That low price is due to all the electricity they import from older dirty coal burning generation facilities in Wyoming.

Old, dirty coal plants produce cheap electricity.  Natual gas is very cheap now due to the fracking of gas wells so gas powered plants also produce very cheap electricity.  Hydro can be cheap.  Everything else is more.

I now a guy with a huge commercial BTC mining warehouse that pays less than $0.02/kWh because he buys power in huge steady quantities.  That is the only way to make money in the mining game long term.

The cheapest type energy by far is nuclear power. It is also one of the cleanest forms of energy.

I really do not understand why the "green lobby" does not push nuclear power more

I don't know where you live, but at least in the US, the Dakotas would be a great place for wind.  I've read articles before describing them the "Saudi Arabia of wind" and that they have enough wind that if enough wind turbines were built there, they could generate enough power not just for that area, but to power the entire country.  The catch, of course, for the rest of the country is that there still aren't a ton of people in the Dakotas, excluding the boom of population that's happening now because of the oil boom.

But if you move there and plan on setting one up, I hope you like cold weather.  It gets REAL cold there.

Cold weather is great for mining.  Free cooling.

Probably just ignorance mostly. Lots of people have irrational fears of things they don't understand and associate with death instead of cleaner and more efficient energy.

I really like what I have read about Thorium reactors.  Seems like the way to go.  If the whole industry was not funded by the need for Plutonium then I think there would be a lot more reseach into things like Thorium reactors.  If you have not read about them check it out.

The green lobby seems to be "well educated" on things like global warming and the effects of other energy types

I'm about as much of an environmentalist as one gets, but I must admit I get rather peeved at my fellow environmentalists who hate nuclear power and wind / solar.  There's always something to object to power plants, but why are you really objecting to a solar power plant installation in the desert?  Yes, I saw this happen.  It saddens me greatly, the continual NIMBY attitude that it seems a lot of environmentalists have does nothing but ensure that we'll keep on being powered by fossil fuels. 

I think water is the number one concern in the desert and PV installations do require some water.

I have a buddy who is involved with very large scale PV installations.  He told me the following story about having to solve the following string of issues:

PV system would get dust on it reducing the output. 
Solution: wash them periodically.

The water run off from the washing caused weeds to grow, when they got high enough they shaded the collectors reducing the output.
Solution:  periodically bring in goats to eat the weeds.

The goats somtimes chewed on the wires causing failures reducing output.
Solution:  they were noodling that when he talked to me.

And so it goes....

I think most environmental want more solar and wind power to be used. This is why they try to get regulations enacted that make it more expensive to produce "carbon" power so that wind and (especially) solar energy can better compete 

This recent coindesk article on is interesting:

http://www.coindesk.com/microscope-economic-environmental-costs-bitcoin-mining/

I think the author is conservative, and the likely power consumption is way higher.


When you have projects like MegaBigPower wanting to add 50PH/s to the nethash each month, that's a new power station each year just for them.

http://www.coindesk.com/megabigpower-launches-global-franchisee-network-add-50phs-per-month-bitcoin-network/


At some point the exponential growth has to stop, if only by governments cracking down on the large scale energy wastage.

Mining uses an insane amount of energy.  I honestly hope governments do crack down on people doing this (at least those who aren't powering it with their own PV setup).  This really is the biggest problem with bitcoin in my opinion, it's a huge energy hog.  I think most of the things about bitcoin taking over would be great, but this is definitely not one of those aspects of bitcoin that are a positive.  (Yeah, I know, the fiat system also uses a tremendous amount of energy, but we really should be striving to do better than them, rather than merely not as bad.)

The whole "mining is a huge waste of electricity" motion is just completely wrong.

Mining currently uses between 105MW and 150MW.
However, the act of mining is the act of keeping the Bitcoin payment system alive and secure.

Do you know how much electricity all the banks in the world use for their "services" to the people?
Hint: It is most likely much more.

Regardless of the above, many miners use renewable energy sources for mining, as in most areas it is simply cheaper.

The example of WA state with 2-4ct/kWh for commercial use is just one example, but it clearly shows that unsubsidized renewable energy can be far cheaper than any coal / gas plant.

Many miners also get hosting in Iceland, having the benefit of 100% renewable energy there plus extremely good conditions for free cooling (my miner has an intake temperature of 18°C in the summer).

Nonsense, the comparison between mining and the banking network is specious. You don't have to mine fiat the power consumption of the banking system is to provide the layer above the currency, the accounting, recording, lending, usury, and other services. Bitcoin still needs those infrastructures as well, it's just that it's not widely accepted so that part of the infrastructure is currently quite small, but growing every day. If Bitcoin becomes mainstream and adopted by banks and credit card providers, you wait and see how much total power is used then!

Mining uses a miniscule amount of energy compared to all the wasteful use of electricity in the United States alone.

How come the electricity consumption per capita is more than 2.5* that of Germany?

If everyone of the 318m US citizens would reduce power consumption by only 10% (easily achieved by switching lightbulbs...), it would save 25 700 MW of generation capacity alone.

You can do all the bitcoin mining of your hearts´content with that, and then some.

And how long do you think it will be before the bitcoin mining network is up to many gigawatts of power consumption, with nethash growing by 60% each month?

If only MegaBigPower keeps up with their 50PH/s per mth. expansion plans, that will require a new nuclear power station every 18months.

http://media.coindesk.com/2014/06/20140523_164302-e1404154208253.jpg

http://www.coindesk.com/megabigpower-launches-global-franchisee-network-add-50phs-per-month-bitcoin-network/

In a first stage they want to build 10MW of capacity until 2015 and are planning to build one up to 50MW facility.

You probably might not know this but assuming exponential difficulty growth is obviously a fallacy (aslong as there isn´t exponential price increase in Btc).

There is a very real limit to difficulty, and it is Btc price, power efficiency, power price, available hosting facilities, hardware price and some other factors.


Regardless, compared to my example, Difficulty could increase atleast 171 fold before being the same amount that can easily be saved by some energy efficient appliances in the household in the US alone.
It really is a TINY usage compared to wasteful electricity usage in the US, let alone globally.

can energy be redirected to do useful things? maybe i'm speaking bullshit i know lol

Its mostly waste heat, so that can have uses.

This picture is mind blowing. Wow, bitcoin mining has come a long way. It's almost ridiculous at this point. How much further are we going to see this go?

The upper limit is calculated in this thread:

https://bitcointalk.org/index.php?topic=281279.0

Based on $100 per BTC and $0.10 per kWh the electrical cost break even point is about 179.7 PH/s.

Adjusting for $600 per BTC (x6) and $0.02 per kWh (x5) would give us 179.7 * 6 * 5 = 5,391 PH/s.

The same thread assumes 0.8 J/GH so:

    5,391 PH/s = 5.391 * 10¹⁸ H/s
    0.8 J/GH = 0.8 * 10^-9 J/H

    (5.391 * 10¹⁸ H/s) * (0.8 * 10^-9 J/H)
 
    = ( 5.391 * 0.8 ) * ( 10¹⁸ * 10^-9 )  ( H/s * J/H )

    = 4.3128 * 10^(18-9)  (J/s)

    = 4.3128 GW

World power production from:

http://en.wikipedia.org/wiki/Electricity_generation

is about 2311.4 GW.

So Bitcoin would use about 0.1866 % of world electrical output in this maximum build out scenario.

BTW in the above scenario the difficulty would be 753,000,000,000

This statement is just very shortsighted. Minuscule is only bitcoin's market cap compared to the world's whole economy.


That is today, at 600$/BTC. Let's assume 6000$/BTC and you'll already be using 1.866% of the world's elecricity use but still be a minuscule economy. (directly following from the calculation above)

This by the way also means hidden cost of the bitcoin network paid by everybody because of increasing electricity prices due to increased demand.

(http://www.pcmag.com/article2/0,2817,2392654,00.asp) also (http://www.popsci.com/science/article/2011-09/google-releases-its-energy-consumption-numbers-revealing-260-million-watt-continuous-suck)

( 2.26 * 10¹² Wh/y ) / (24*365.25 h/y) = 257 MW

(http://www.techthefuture.com/technology/how-much-electricity-does-the-internet-use/)

Current minimum consumption of the Bitcoin network (it is probably more than this due to AC needs and the use of less efficient older hashing units):

From https://blockchain.info/charts/hash-rate lets call it 120,000,000 GH/s

(1.2 * 10¹⁷ H/s) * (0.8 * 10^-9 J/H)
 
    = ( 1.2 * 0.8 ) * ( 10¹⁷ * 10^-9 )  ( H/s * J/H )

    = 0.96 * 10^(17-9)  (J/s)

    = 96 MW

Let's double it.  Still less than Google at the current time.

We should surpase Google for sure.

I have just updated my thread here:

https://bitcointalk.org/index.php?topic=518111.msg7626097#msg7626097

where I show what would happen if the growth in hash rate over the last one year period (23.45% per adjustment) were to continue at the same rate for one more year.

It shows that something (price or growth in hash rate) must give before the end of the year.

You can also see the current power and energy numbers and most importantly for this thread the projected energy and power consumption if the growth in hash rate were to be maintained for just one more year.  Don't worry it can't and it won't.

The guy is clearly wrong, he claims the Bitcoin emission trend is decreasing. Then he quotes the nethash as 95PH/s yet it's already hit 145PH/s a week ago. Trying to compare Bitcoin mining with physical mining in the real world is obviously a stunt and an attempt to distract and uninformed reader. There is no relationship between the two processes other than they both use the word "mining'.

What he fails to grasp is the exponential growth in bitcoin mining power consumption.

I think one way to slow the growth is reduce the block reward from 25 coins to just 1 coin per block, there is no way the BTC price will jump 25 fold to compensate as more then half the coins are already mined so there is no shortage of supply, and no reason for that much fiat to enter the market.

Your scenario is not very likely to happen.
And good luck with network security if mining power is fluctuating heavily.
Of course, if miners would be distributed equally around the globe, electricity spot prices would even out quite well and so would actual hashpower. But a heavy centralization of mining in the countries with lowest electricity prices is much more probable in your scenario.

I don't see how your suggestions could make things better. The only way out I see is a hard fork to PoS. With today's market cap that should be pretty secure. Or all of us support energy-sufficient altcoins by holding more value in a PoS altcoin than in Bitcoin!

According to this formula:
your suggestion might make things 25x better. But there would be a shortage of liquidity on the exchanges driving BTC price. So it's not 25x.
And it only works if people do not pay more transaction fees.

Great minds think alike (you and Satoshi)!  In fact, the block reward will be reduced to just 0.78125000 BTC per block in about 2033. 

See entire schedule here:

More than half of all the 21million bitcoins that were ever designed to exist are already out there in circulation. Mining new coins is not important to liquidity, we just need to find 144 blocks each day for transactions to ride on, it no longer requires many coins in those blocks as we have enough in circulation, so it might as well just be the one coin. I don't think waiting until 2029 is the answer, the world will be struggling for oil and coal around then unless something else happens to curb world demand.

The theory is that there will be enough TX volume that TX fees make up for the lower block subsidies over time

Oh.

You are suggesting a change to the protocol.  Not just any change but a change that would cause a hard fork.  And, not just any hard fork but a hard fork that would greatly punish the miners that switch to your suggested alternate protocol (only 1 BTC/block) versus staying on the original protocol (25 BTC/block).

So, it will never happen.  And just so you know when I said never, I meant never and there is no need to try to argue for the various virtues of the proposal - unless you actually like talking to a walls or trying to move mountains.

Yes, of course. 

BTW have you (or anyone else) seen a good technical thread that tries to estimate the fees over time.  For a project I am doing I need to combine the chart I just published with a sound estimate as to how the fees are going to grow over time.  Just wondering if anyone has spotted such a thread anywhere.

I am quite sure there will be more BTC hard forks with protocol changes.  I have read some of the proposals. I don't think there has been much discussion on energy saving though.

There are dozens and dozens of theads discussing the energy consumption of the Bitcoin network and proposals on how to "fix" it - mostly by going from POW to POS but there are some others to be found with some searching.

Just search for "Bitcoin" and "POS" and you will find some hard fork proposals that might interest you but, of course, do not interest any of the miners who just bought ASICs.

PoS has it's problems, it's currently on the prohibited changes list.  https://en.bitcoin.it/wiki/Prohibited_changes

PoS would simply make Bitcoin almost centralized since so few control so high of a percentage of all the coins

PoS would simply make Bitcoin almost centralized since so few control so high of a percentage of all the coins
[/quote]
So we don't have to discuss that as a possible option/solution.  Cool.

I have the same concern with PoW mining, so I don't see the difference

Technology stills evolving, more efficient mining hardware will appears, adding that not only Bitcoin has energy consumption problems, the rest of the world needs now alternative clean forms of energy that i believe are being researched so hard right now...

Besides, current prices are too low, some things would be different if Bitcoin goes to the thousand again...

The problem would be worse if the bitcoin price rises, as mega farms could afford more energy to waste.

Efficiency of mining hardware has no effect on energy consumption of the network. It only drives difficulty.

This is not true. If more efficient miners are on the network then the average energy consumption per THs on the network will be lower. As the difficulty increases, less efficient miners will take their machine offline as it will be consuming more energy then the bitcoin it produces will pay for.

Yes, efficiency does not matter when calculating the power consumption of the entire Bitcoin network.

Power consumption only depends on five things:  the exchange rate of BTC, the era, the average amount of fees per hour, the cost of energy, and the average gross profit margin of the miners.

Watch this:

x = exchange rate [USD/BTC]
e = era [0..32] (we are currently in era 1)
f = average fees per hour [BTC/hour]
c = cost of energy [USD/kWh]
g = average gross profit margin [unitless ratio]

From the era we can caclulate the average hourly BTC subsidy rate:

s = 6(50/2^e) [BTC/hour]

And the amount of BTC all the miners in the world would make per hour:

b = s + f [BTC/hour]

From this we can calculate the amount of USD per hour all the miners in the world would make:

u = b(x) [USD/hour]

Given the worldwide average gross profit margin the amount spent worldwide on energy would be:

u(1 - g) [USD/hour]

And finally, the worldwide power consumption would be:

P = u(1 - g)/c [kW]

  = b(x)(1 - g)/c [kW]
  = (s + f)(x)(1 - g)/c [kW]
  = (6(50/2^e) + f)(x)(1 - g)/c [kW]

Notice that efficiency does not enter into this equation and does not matter.

You do not need to know or estimate the average overall efficiency of the mining network unless you want to calculate the difficulty and/or hashrate.

I would disagree with your statement about there being no such thing as "efficient" mining hardware. I would define efficient miners as machines that use less electricity per GHs then the network does as a whole.

Your theory doesn't match the facts. People just keep adding more mining hardware until the electricity consumption is as much as they can stand. The diff rises and they get less BTC than they were getting with the lower efficiency setup.

Of course, mining hardware can become more efficient.  The point is that does not matter with respect to the total power and/or energy consumption of the Bitcoin network.  Look at the math above where I proved this to you.  

https://bitcointalk.org/index.php?topic=520977.msg7687843#msg7687843

Also, if that is too much to read then re-read this until you understand it:


See yet?

In theory, you could add a little bit of additional mining capacity (that is very efficient in terms of electric usage) to the network that would increase the difficulty a small amount, but that small amount would make it so a lot of less efficient miners no longer mine enough to earn a profit after electric costs so they are taken offline, and then subsequently replaced with more efficient miners with equal hashpower but with less overall electric consumption.   

Other payment methods use electricity as well. It is just that it is not as transparent as to how much electricity visa for example uses to process their payments. Visa must power their office buildings, computers for employees, computers that process transactions, banks that send money to/from visa (their employees, offices, and computers), printers that print customer statements, call centers to resolve disputes.

Don't confuse mining crypto with payment networks, the two are not the same. Fiat doesn't need mining energy.

The amount of energy used for the Bitcoin Network is tiny compared to the amount of money to print, mint, guard, secure, etc for say fiat currency.

For now, perhaps.  But how big would this energy consumption be if the bitcoin network were as big as a traditional major finanical company like, say, Visa?  We should strive to do better than fiat is doing, not merely being not as bad as the traditional system.

Visa doesn't mine coins, their network has nothing to do with the creation of money, the banks do that for them. Your comparison is flawed.

If not them, then who would you compare bitcoin to?  I grant you that it's not a perfect comparison, but there is no one to one comparison between bitcoin and the traditional financial world: the bitcoin network / ledger / mining combined creates new coins, processes transactions, handles transaction fees, etc. 

I picked a major player in the traditional financial system to choose to compare.  But if you want to say that it's a flawed comparison, then you're missing the point here.  The point is if you're going to make a comparison, you have to compare bitcoin to something.  And bitcoin does what credit card companies (like Visa) do, it does what Banks do in creating money, it handles transaction fees, which traditionally would be something credit card companies do.  So...what would you compare bitcoin with?  If you want a true one to one comparison, you're shit out of luck, because there isn't any one to one comparison since the bitcoin network does what several different players of the traditional financial system do. 

The comparison flawed, it's like trying to compare cars to roads, it just doesn't make sense. The Visa style payment gateways sit on top of the currencies. Visa could easily add Bitcoin as one of the supported currencies. Visa chooses to use fiat atm, but I can see it adding some crypto in the future, and it's infrastructure or energy consumption will hardly need to change as it wont be mining the crypto.

I sometimes think Satoshi was pushed by Energy lobby's.

I really don't see credit card processors using crypto currencies ever. Bitcoin was designed to make it cheaper then using something like a credit card. If someone were to use a credit card to transact with bitcoin then they would lose the advantages of using bitcoin in the first place.

They might, if we ever get to the point (many years into the future) where fiat becomes obsolete.  The advantage of credit cards is, at least in part, for people living paycheck to paycheck, they can spend money that they don't have yet and pay the bill later when they've been paid.  But I agree with your overall point, it does seem rather counter-intuitive that someone would use both a credit card and bitcoins at the same time.  Debit cards would make more sense, though.

I have to imagine energy consumption is through the roof.

That is an interesting point with respect to the dream of having 1 BTC = $100,000 (or pick your favorite high number).

Using my previously derived formula for the power consumption:

P = (6(50/2^e) + f)(x)(1 - g)/c [kW]

where:

x = exchange rate [USD/BTC]
e = era [0..32] (we are currently in era 1)
f = average fees per hour [BTC/hour]
c = cost of energy [USD/kWh]
g = average gross profit margin [unitless ratio]

we can look at the power consumption in each era assuming a price of $100,000 per BTC.

In order to make it simple I will make the following assumptions:

x = $100,000 per BTC
f = fees per hour will keep the coinbase above 6 BTC/hour (1 BTC/block) in all eras
c = $0.10 per kWh
g = 0.1 miner gross profit margin

Here is my entire spreadsheet, I was too lazy to format it last night:

Not really.  It's more proportional to the number of miners because the electrical use is pretty much all from mining.  Of course, greed being what it is, the higher the price of bitcoin goes, the more people are likely to mine, so in a correlational sense, you're probably right, but it's definitely not cause and effect.

I personally think you have some flaws in your analysis.  One of them is the fairly mythical concept of the rational person.  One example of this is sunk costs, you're right that a rational minor should ignore the cost of the mining rig.  I do my best (and probably a fairly good job) of applying the concept of sunk costs to my own life.  But a lot of people don't, even people who should know better.  So I suspect that a lot of miners continue to mine even if it's unprofitable with their miner from an electricity standpoint simply because they've seen the huge runup in price and hope that this trend will continue over time.

Granted, you did point out these caveats in your own reply to your post.  I just feel that they should be more prominently in the actual analysis because I strongly suspect that a lot of miners are speculating on an increase in price so they continue to mine, whether or not it's profitable.  I can understand the idea, I'm speculating too, but I did it just by buying bitcoins rather than spending money on mining gear.  Mostly because as much as I've been tempted from time to time, I really think at this point, mining is a bit of a zero sum game by now.

I think comparative advantage is applicable to that one.   It might be expensive and pointless to some but some parts of the world have much lower electricity costs.  This can be seen in aluminium processing which favours local low electric costs but many parts of the world this business is not profitable afaik.

Bitcoin has the advantage of being highly transportable work so more then anything this rules applies.    It would make sense for those stuck with old bad equipment to maybe sell it in bulk to the low cost parts of the world.

Anyhow I think moores law will reduce electric cost over time until the network consumes no more then a LED readout on any device.  On average of course and I get that difficulty must keep rising and so people assume higher electric cost but I dont think its about electric it'll be about technology cost and in the end R&D and maybe silcon production will be the networks processing cost more then other factors

You're quite right that electricity is cheaper in some places in than others.  I only hope for the environment's sake that the bitcoin miners are gravitating towards places like Iceland which has plentiful renewable power rather than places that are cheap simply because the electricity is subsidized.  I don't buy your comment about Moore's law though, bitcoin mining is an arms race for getting the best gear the fastest.  While that gear may become more efficient eventually, I'd be incredibly surprised if it ever uses that little power. 

In any case, Moore's law typically applies to processor speed, not power usage.  Also, the other reason I don't buy it is that Moore's law is treated by some like it's an immutable thing that's set in stone.  It's not, scientists very close to reaching the physical limits of how much they can cram into one chip.  So I'd be surprised if Moore's law continues into the future very long like it has in the past thirty years.

Debit cards would make even less sense then credit cards as they have the same costs as credit cards but would not have your counter point as above.

I doubt that any crypto would ever be so widely used that fiat would become obsolete as it is near impossible to garnish bitcoin from someone. Today for example if you were to spend $5,000 with your credit card, then simply don't pay your credit card bill despite having the ability to do so and have say $10,000 in your bank account then the credit card company can get a judgment against you and the court can order your bank to give the credit card company the $5,000 that you owe out of your account via a garnishment. This would simply not be possible with bitcoin.

All the posts about efficiency and Moore's law and power consumption per device are totally off base.   Also your comment about people who mine when it is not profitable is really off base because that behavior makes things even worse as far as the total power consumed by the network goes.

The amount of power consumed is proportional to the amount of available income in the mining process.  More efficient mining hardware just leads to more miners and a higher difficulty, it does not affect the total power consumption.

I was looking for an analysis that's something like this. You helped me save hours of searching through the internet.
Back to the topic, if more and more energy efficient miners come out, then the energy consumption of the whole network would decrease. But if widespread adoption of bitcoin ever happens, then the energy consumption would again plummet up and may become an issue because it may cause the global consumption of energy to rise up too. Great analysis though, thanks for helping me understand the maths here. :D

You are missing something fundamental about the mining here.  If bitcoin is worth X in 2020 that means miners will spend 12.5X on electricity every 10 minutes in 2020.  It will be worth it for miners to spend that much on power; that's what we mean by "worth X".  No technological improvements can change that, it is the fundamental algo of bitcoin that miners get new coins.  If coins are worth more they will spend more on power.       

You are welcome!

What you said here is not true:


What is true is:

If more and more energy efficient miners come out, then the energy consumption of the whole network would stay the same.  This is because people just buy more and more of these more energy efficient miners.  They will keep buying more equipment and adding to the hashrate and difficulty until they bring the power consumption back up to the maximum level allowed by profit.

So as I have said before many times efficiency does not enter into this equation:

P = (6(50/2^e) + f)(x)(1 - g)/c [kW]

only the following parameters:

x = exchange rate [USD/BTC]
e = era [0..32] (we are currently in era 1)
f = average fees per hour [BTC/hour]
c = cost of energy [USD/kWh]
g = average gross profit margin [unitless ratio]
 
Notice that price (the exchange rate x) does enter into the equation so what you said here:


This is because widespread adoption would drive the price higher and price does enter into the global energy consumption equation.

Interesting discussion...

RE: Solar and Wind - For every megawatt of solar or wind power added to an ISO there must be a megawatt of NG, coal or nuclear power deployed in order to maintain stability of the grid. ISOs such as ERCOT must have the ability to maintain stable supply / frequency which means if a wind farm is feeding the grid then ERCOT must have adequate firm fixed generation capacity such as NG fired spinning reserves to ramp up as the wind stops blowing.

The only way to avoid the need to provide an equal amount of traditional generation to pair with your wind or solar is if you have the energy storage capacity to create an adequate reserve. Batteries are still cost prohibitive as are most other methods. The other option would be to curtail operations as the wind and sun stop blowing or shining.

RE: Huge Opportunity That Bitcoin Is Perfectly Suited For - When you start to look beyond the traditional constraints of the existing grid and fuel transportation networks you will find that Bitcoin is perhaps one of the best suited technologies to leverage abundant energy resources that are hiding in plain site. Rather than spoil it by elaborating too soon I am curious to know if any of you smart people know what I am referring to. Anyone?

There is no way to transmit the power needed from the sunny side to the dark side.  Sorry.

RE: Solar and Wind very interesting point that the solar/wind people fail to realize or fail to mention when they push build out of these technologies.

RE: Huge Opportunity That Bitcoin Is Perfectly Suited For sorry, not smart enough to guess.

This may or may not be what you are looking for, but one option would be to take advantage of variable electric prices that are present in some places of the country. For example, the price of electricity is cheaper at night when people are sleeping as this is generally an off peak period for electric usage. You could only run your miners during this time if it would be unprofitable to run it during the day.

Another potential option would likely not be something that an individual could do, but rather a public policy that an individual could attempt to change - to try to get more nuclear energy to be provided to the electric grid.

An easier option than these super advanced solar beaming technology would be to have people in the Dakotas build wind turbines and host miners on behalf of other people.  The Dakotas have been described as the Saudi Arabia of wind and have the potential to generate enough power to power the entire United States.  The catch is, of course, that being ridiculously cold much of the year, almost no one lives there and much of the power generated would likely be lost to power line inefficiencies.  But the turbine powered miner technique wouldn't have this problem as the product being sent elsewhere would be bitcoins, not power, which wouldn't require the creation of a whole new system of power lines.

It's definitely a potential market for anyone who lives in this area.  They could even use the miner as a space heater during the winter months. 

Why would nuclear energy not be able to replace other, dirtier forms of electricity? Nuclear energy is near unlimited, very cheap, reliable and clean.

Your other "green" energy sources are unlimited, however are expensive, and unreliable.

Hindenburg fireball isn't caused hydrogen.  Its the coating or paint on the outer skin that burned

China is putting more GE Gen5 reactors online and India is starting phase 3 of it's thorium based long term plan. I think the rest of the world has pretty much given nuclear a kick in the pants.  The US hasn't even approved construction of a single new reactor since the 70s.

All off topic.  This thread is about energy consumption of the Bitcoin Network.  You should go start your own thread discussing how you think energy should or should not be produced.

Thanks.

No worries, we can just let this thread die then.

Sounds good.  We will start another thread about estimating the energy consumption of the Bitcoin network and you can start another thread to discuss energy production.

You can have this thread.

A new thread specifically discussing estimating the power consumption of the Bitcoin network is here:

https://bitcointalk.org/index.php?topic=694401.0

Here is an article from end of last year on the this topic:

The Power Consumption of the Bitcoin Network:
Are we destroying the planet?
https://docs.google.com/document/d/1Lb_iLMmH5sIiYfGeHr9tCTnaoWjK25zDlvO0nXXgEng/edit

6 Months old and already totally outdated though ;-)

Original thread is here (the author is Xyver):
https://bitcointalk.org/index.php?topic=381518.0

That seems correct but work has been done.     Cotton notes are not able to save or give energy either but they can convey work.   That work can be used in exchange for energy elsewhere and btc will do this also.


The second argument is btc work useful or not, its describing secure and unique hash codes.  The world has found some use for btc in security and so energy spent is useful and can be used in one place and then elsewhere the btc can be reused to pay for energy just like notes are.  
Im not saying btc has to be valued by all and Im not sure any energy company takes btc as direct payment, probably not but in effect btc is conveying energy or if you prefer, work


It does seem viable that places with excess energy will be connected to the internet more easily then they can transport large amounts of electricity ?   If that is the case, btc is being useful to convey work/energy usage ?

The block subsidy will be reduced to under 1BTC.  Great minds, yours and Satoshis, do think alike. 

I was just wondering how much power the bitcoin network was consuming and ran some numbers myself:

Let's assume everyone mining bitcoin was running Antminer S3s or at least a equally efficient rig(which we all know isn't the case) and just to make things a little easier well say the S3 runs at 500Ghs and consumes 350W

Right now were up to 214PHs on bitcoin so thats 428000 of these S3s consuming 350W each so 149,800,000 Watts

Now I found this:  http://www.eia.gov/tools/faqs/faq.cfm?id=104&t=3

"The Palo Verde plant in Arizona has three reactors with the largest combined generating capacity1 of about 3,937 megawatts (MW)."

Which I assume comes out to 3,937,000 Watts

So you would need 38 of these high output Nuclear Power Plants to power the Bitcoin network.  That is fucking mind boggling! 

One more fix needed I think ;)

But that just happened last two weeks. It will keep on rising in the (near) future.

We can look at the (necessity of building more) nuclear plants as positive externality of Bitcoin. :)
Might even stimulate Fusion research.

If only we could conduct the heat generated by big ass miners into tubes that would deliver free calefaction for people, bitcoin would be something even greater.

Here's my take.

The Math: (extract from an internal report)

• First difficult y drop seen in 2 years occurred 2 Dec 2014
• Over the next 85 weeks (21 months) a total of 2.1 million Bitcoins will be mined
• USD $735 million at $350/BTC, $1 billion at $500/BTC, >$2 billion at $1000/BTC
• 280 MW estimated Bitcoin network size today (compared to Google's estimated 260 MW, Facebook 78 MW)
• Block reward halving ETA July 2016 (85 weeks, 5 days as of today)
• Bitcoin output from 25 BTC to 12.5 BTC per block (next halving estimated 2020)


Historical Bitcoin Network Data:

Nov/Dec 2014 (today):

Hashrate:                     290 PH/s    (+35%)
Total Megawatt:                 280 MW   (home mining almost gone)
Megawatt Industrial/Data Center:    270 MW   (+35% 0.85W/GH new capacity)

August 2014:

Hashrate:                     214 PH/s    (+375% or 94%/month)
Total Megawatt:                 214 MW
Megawatt Industrial/Data Center:    200 MW    (>1000-2000%)

April 2014:
Hashrate:                     57 PH/s
Total Megawatt:           250-500 MW    (mostly at home miners)
Megawatt Industrial/Data Center:    10-20 MW



Bitcoiners are in fact the first ones to heat complete greenhouses.

AsicMiner testing an 0.25 W/GH board.

https://bitcointalk.org/index.php?topic=888260.msg9794133#msg9794133

If it is implemented in miner, that would be interesting.

This is a good alternative energy that can be used to support the massive electricity consumption


What a great information here, gonna note this !

Interesting. Note taken.  ;)

they would rise the price of it, until it won't roi for sure, just like they did with their previous stuff

but it's cool to see something x3-4 more efficient

This appears to still be in the very early testing phase and will probably not be available for some time. Also more efficient miners will only result in overall higher hashrate on the network which means a higher difficulty, not a lower amount of electricity used

Many megawatts I have personal knowledge of (air cooled) have been switched off recently. If this will be "compensated" by newer hardware coming online (hardware that was already in the pipeline) remains to be seen short term.

Probably not (although difficulty may decline some).   What is more likely is a rotation.  Older obsolete hardware getting scrapped and higher electrical cost miners being replaced by lower electrical cost miners.   The boom prices lead to some pretty weak miners.  There were miners setting up shop in UK with $0.25+ per kWh.  The OP talks about an average price of $0.15 but I think the upper end is going to be closer to $0.10 and the average somewhere around $0.05 per kWh.

why they can't redicrect the energy dissipated, to something more useful, right cure the cancer or something

Why can't a bank guard also cure cancer at the same time?   Not to be snarky but POW has certain important properties:
1) it needs to be cryptographically related to the current and prior blocks (which is why we hash the blockheader)
2) it needs to be a probabilistic process (any hash can solve a block but on average it will take a huge number to do so)
3) the difficulty of the solution needs to be adjustable
4) the solution on average needs to take a large amount of work to solve but take a very small amount of work to verify
5) a potential solution should be reached by any person with only access to the blockchain and memory pool (no central authority)

This is a very difficult set of requirements which is why POW is the "magic" that made Bitcoin possible.   ecurrency is various forms existed prior to Bitcoin but they required a central authority as the idea of there was no solution on how to achieve a distributed consensus among untrusted peers.

Still the heat can be used for additional work.   I think as ASICs become more commodity based (i.e. when I can buy a reel of them for immediate shipping at digikey) and energy costs dominate we will see integration of ASICs into systems that heat a house or hot water tank with the waste heat.

I repost what stdset replied on https://bitcointalk.org/index.php?topic=920007.60 (https://bitcointalk.org/index.php?topic=920007.60)


I have checked his model with today's input, and it match with reality. I have not seen any flaw in his reasoning.
My question was : why do you think POW energy consumption to be unsustainable ? It was a very good response I think.
However, we won't reach the problem until bitcoin cost 1 Million $ in 12 years ;)
Even if "6% of total electric power is a bit much" is highly subjective depending on the utility of bitcoin. Also the heat can be reused.

The input he took are conservatives, making them more realistic cost more energy.

I seriously doubt that "reusing heat from miners" has any sound foundation, I think that cooling miners is just an extra cost. Heat extraction and concentration may well be uneconomical in most cases, Also converting ac to dc adds to inefficiency.
It could well be that having solar powered miners in sahara even with the added cooling costs may be preferable than expending electricity in say norway and tapping into byproducted heat

Well, if we agree on the fact that if bitcoin rises to 1million USD in 12 years, it will consume 6% of power (according to the model I posted from stdset), the question becomes : will this provoke an incitation to move to POS cryptos ?
I am not sure about that, because if it rises to 1million, the cost of migrating to something else are astronomical and might outweigh the cost of energy.
However, if the price of Bitcoin would make some business model impossible, I would expect migration to happen.

We can slow down the process with energy reuse, but it does not solve the question, only post bone it.

Last century:  Cost of production of money: 0 watts.   Massive waste of all resources results.

This century:  Cost of production of money: determined by public currency mining market.  Efficient use of resources results. 

Yep. Considering the OP still contains this major factual error.

Not so
It is not 2jobs it is vertical integration
And quite safe as investment as you diversify and get flexible as to what to produce bitcoins or electricity depending on where you get more value.
Iwould first deploy solar panels and on a second phase the miners and iterate reinvestment on around robin fashion.

A note on heat extraction : powerplants can reuse heat because they produce highly concentrated heat, miners produce heat of low value.
A note on green energy, best use energy sources that provide dc so as to skip conversion and related loses, solars are the best type imho to drive circuitry

Thiss so hard. The amount of effort put into generating Bitcoin is justified, we are talking being backed by math there boys. The result will be an harmonic one, and not a catastrophe (fiat).

Cost of production of money is 0, but the money transmission is huge due to the cost involved in running the banks and data centres.

If something costs 0, you have right there a big reason to think it's a scam.

The cost of mining is equal to the fees incurred. When the value rises, the percentage of fee will drop. I agree the total cost of mining will rises but not linear to the value of BTC. The above comments does not apply to the initial mining period when new coins are issued.

Well since the subsidy will not go to zero for another 120-130 years most discussions about costs, energy, and sustainability deal with the period from now until ~2140.   Even after the next couple halvings the subsidy will play a major impact in the cost of the network.