Bitcoin Forum

2013 was a year where the network hash rate grew over 40,000%.   As I write this the network hash rate is 16,371 TH/s.     If you assume 600W per 1 THs which is probably conservative, it works out to about 10 MWh (Mega watts hours) being spent on bitcoin mining.    This is still a tiny fraction of the world’s total electrical power production which is currently around 22,500 TWh.  (Terra watts hours)   Source: http://yearbook.enerdata.net/#world-electricity-production-map-graph-and-data.html

So far no concerns but look what happens if the growth continues at the same torrid rate:

World's Electrical Power Production in 2012 ~ 22500 TWH
Current power consumption of bitcoin hashing ~ 600J/TH/s * 16371 TH/s = 9,822,600W

(Going with BurtW correction here)

Assuming same torrid growth rate of network growth and constant power production
NOW     : 85.6 GWh        -  0.00038 % of World's electical power production
Year +1 : 34.2 TWh        - 0.00152 %
Year +2 : 13696 TWh     - .6 %
Year +3 : 5478400 TWh  -  243.48 % of the World's electircal power production

So it seems likely that soon the network growth is going to start hitting serious limits.   Keep in mind I’m only looking at electrical power here.   There are many more potential limits like just running out places to house the mining and producing smaller faster ASICs will become very hard.

Thoughts?

Edit note - Many corrections made ... thanks!  

The electricity cost in mining will not be the major cointributing factor in foreseeable future

In GPU era, you can buy GPU with instant delivery anywhere in the world, but with ASICs, you have to buy from another country or wait for the pre-order to arrive, most of the cost has been shifted to mining equipments

That cost will not go down due to competition, since there is a consensus for mining equipment makers to follow the others in their pricing. Most of those companies on the market sell at a price at bitcoin's tradition of 3 month ROI.

Currently the difficulty is rising fast, so they have to cut the price aggresively, but if it slows down, it will be much easier for miners to reach ROI in 3 months, so they will also stop cutting the price

And another issue is that chip makers do not have enough capacity for producing enough bitcoin hashing chips

No, the burned electricity does not contribute to bitcoin's value in any meaningful way. Bitcoin is an asset with some rules about supply and a market, that is why there is a price. Enforcing those rules which make it valuable happens to require power.

By your understanding, what happens if bitcoin suddenly switched to pure proof-of-stake where the computational cost required to obtain distributed consensus was close to $0?

I do feel like a villain from the series captain planet...

Do you remember Alpha Centauri?  :)

Resources exist to be consumed. And consumed they will be, if not by this generation then by some future. By what right does this forgotten future seek to deny us our birthright? None I say! Let us take what is ours, chew and eat our fill.
--- CEO Nwabudike Morgan "The Ethics of Greed"


Our first challenge is to create an entire economic infrastructure, from top to bottom, out of whole cloth. No gradual evolution from previous economic systems is possible, because there is no previous economic system. Each interdependent piece must be materialized simultaneously and in perfect working order; otherwise the system will crash out before it ever gets off the ground.
--- CEO Nwabudike Morgan, "The Centauri Monopoly"


And when at last it is time for the transition from megacorporation to planetary government, from entrepreneur to emperor, it is then that the true genius of our strategy shall become apparent, for energy is the lifeblood of this society and when the chips are down he who controls the energy supply controls Planet. In former times the energy monopoly was called "The Power Company"; we intend to give this name an entirely new meaning.
--- CEO Nwabudike Morgan "The Centauri Monopoly"

Wait.  Is that a good or bad thing?

Please fix the OP so I can read it.  I cannot read any post with W/h in it as it caused my little EE brain to freeze up.  See the second post for an explaination of units, fix the OP, then I will read it and comment.

10 MWh is 10 megawatt-hours.

The amount of energy that it takes to transfer 1 BTC from one address to another is a small fraction of the energy (and none of the paper) required to transfer the equivalent amount ($800) from one bank account to another. Bitcoin mining is saving the world's resources, if anything.

Everyone seems to think the market cap of the coins in a network matter... I mean, I guess it does to a degree, but the value of bitcoins as a concept is the payment network - not the coins themselves.

Sorry for the confusion, I'm just trying to keep it simple.

Wh commonly refers to 1 Watt per hour which what most people understand because they pay for electrical power by the kWh.   If I translated to Joules most people wouldn't understand.   TWh is what is often being used  in many sources of information, for example: http://yearbook.enerdata.net/#world-electricity-production-map-graph-and-data.html


I don't follow your logic.   The amount of power consumption used to produce a bitcoin isn't the same as the value of the bitcoin. 

kWh = kilowatt-hour != kilowatt/per

Just write out your units for everything all the way down to kg, meters, and seconds and you'll see

Actually I used 600W for 1 THs which is basically the power used by ASICs.    You are correct that at this moment power isn't a limiting factor this year.   Next year that will no longer be true as the total network hashing starts consuming 1% of the world's total power production.

Personally I think this is a good thing.   The rate of the hashing growth will slow anyway unless there is a new breakthrough.   After 20 nm ASICs, it will get much harder to reduce power while increasing speed.

I agree, you did a better job of expressing it!

The cost of producing a bitcoin tends to approach the value of a bitcoin because if it costs less than a bitcoin then more people will mine, raising the difficulty and thus the cost. If it costs more than a bitcoin then less people will mine, wihich will lower the difficulty and thus the cost.

I realize that, I'm only using "units" that people commonly see when they try to look up the information.   "TWh" is being used all over the web to talk about energy production and consumption.   The notation is just to make comparisons simple for people and when you are talking in terms of the total amount of the world's electrical production consumed, the units all cancel out.    

As we keep telling you "TWh" is terawatt hours it is NOT terawatts per hour!  Terawatts per hour is WRONG.

Got it now?

Now, go fix your darn OP.

So you are saying it is the goal of bitcoin to consume most of the world's power?

OK now that you got that fixed we can work on the rest of the mistakes in your post:

Hash rate is measured in "hashes per second" = H/s or "terahashes per second" TH/s not THs.

Fixing your units the network hash rate is 16,271 TH/s (terahashes per second)

You are using 600 W per TH/s so: 16,271 TH/s times 600 W per TH/s

This equals 16,371 * 600 = 9,762,600 Watts

Ok you rounded up to 10 MW, fair enough.

That is POWER.

To get units of ENERGY so you can compare it to an annual ENERGY production of 22,500 TWh you need the number of hours in a year:

   24 hours/day * 365.24 days/year = 8765.76 hours per year.

Annual ENERGY consumption of the Bitcoin hashing would be:

   9,762,600 Watts * 8765.76 hours per year = 85,576,608,576 Wh = 85.6 GWh

so 85.6 GWh / 22,500,000 GWh = 0.00038 %

Now the simple answer to your post:

You said "2013 was a year where the network hash rate grew over 40,000%." and then assumed this rate of growth would continue.

It will not.

That is actually my point, but everyone seems to assume that it will.   Your correction ... makes the case even more clear.   

With a great deal of help from the community here, we ended up with numbers that likely reasonable.   While technology will probably improve the power consumption some from my current conservative estimate of 600W per 1 TH/s, it probably won't improve by a factor of 10.   

Most of the world's electrical power is used for basic needs.   Almost 3 years ago Tokyo lost about 10% of it's power due to the problems at Fukushima after the tsunami.  There were rolling "planned" blackouts in parts of the city for months, elevators, escalators and fountains had to be shutdown, trains were put on slower schedules and everywhere you would see signs that we shut this down to conserve power.   Tokyo is probably one of the few cities in the world that could handle an impact like that.   Now almost everything is back to normal because people have managed to conserve the power that was lost.

As was pointed out the global network can't keep growing at this rate, although from the aspect of power we might have more than a year of really rapid expansion.   The power limit is much closer than I though it was at first, but it approaching fast.   This is interesting.   

As I pointed out before, the cost of mining will approach the value of the mined bitcoins.

At $800 per bitcoin, that is currently about $1 billion per year and about 13 TWh per year (at $0.08/kWh). As the block reward drops, the amount of energy used would also decrease (unless energy gets much cheaper), except that the price of bitcoin will probably rise much more quickly.

If the price reaches $40,000 this year then the energy usage could be as high as 650 TWh, assuming there are no other limits.

Thanks for the clarification of your statement.   I can see what you are driving at.   Thankfully, the actual cost is much more complex than just power, but I have to agree that the demand for profit will drive up the network hash rate until it approaches the current value of bitcoin.

This is a very strong argument for just buying and holding bitcoins.

Sure that would be a major jump but it would only take you so far.   I think what might happen is that power rates where there is excessive mining will increase.   The network growth will switch to a sustainable level because of economics or regulation or maybe even wars.   

The amazing thing to me is the extreme growth rate isn't good for anyone but the suppliers right now.   Yet so many don't seem to accept that it can't continue with such an extreme growth rate.   It is common for extreme growth to flatten out to sustainable levels.   The same will happen the bitcoin network.   When it does it will be a good thing for everyone except the suppliers.   But at least they got the rapid money up front.   

The network is growing an exponential rate largely due to huge advances by moving to ASICs.  Producing more of the same ASICs will allow for more of a linear growth rate which could be sustained for a long period of time.   You don't get the same exponential jump in ability without a new technology.   That could happen, but it is unlikely to happen in the near future.   

The current ASIC companies are living high on the hog even demanding and getting pre-payment.   Most of those companies won't even have the ability to cost reduce down to the marginal cost of production.  Additionally why would they?  What manufacture sells goods without a reasonable profit?   So that would probably be another short term limiting factor.  Plus if everyone started using these miners it would have a massive impact on power consumption and probably result in higher power costs, which people wouldn't like.


That is only true from the standpoint of a current miner.   What will happen if the network keeps growing at the insane rate it is, there will be increased power costs and/or regulation to stop it.   It isn't reasonable to think that is would be allowed to grow to 10% or more of the world's electrical power.   Do you think people will be willing to live with vastly higher power bills and rolling blacks outs to support bitcoin mining?    In the CPU/GPU era you weren't using nearly as much power as in being used today. 

What is more likely is that the cost of entry will get so high that it will block most from mining at least bitcoin.   If the cost of power is truly what limits bitcoin, then bitcoin actually won't have much of a future.   Most businesses run with as high as profit margins as they can.   The micro-economics don't work the same on the marco scale and the bitcoin network is way past the micro side. 

Not likely, since Bitcoin is already in a huge bubble.

If it costs $15 in electricity to create $800, capacity will be found.

ASIC producers had all their development costs paid for by the pre-orders, and once you can mass produce a $10 chip that uses $300 worth of electricity, nobody can stop those that can create chips to mine for themselves.

Same as the Bitcoin cost when mined, and its price on the exchanges,  the cost to produce mining hardware and the price they can sell them for, have nothing to do with each other. 

Well when I use to watch the show as I kid, I never really understood the villain's motives - they seemed to just want to pollute for the sake of polluting.  I mean, clearly there is more utility in BTC then there is pollution for pollution's sake, but, I would have to say, bad thing still.

I am looking forward to this time of BTC crowfunded fusion nuclear plants.  :)

As pointed out before, the growth in energy usage by miners is limited by the value of the bitcoins received by miners. At a value of $40,000, the energy consumed is around 650 TWh, which by your calculation is only 0.003% of the total. 10% implies a value of around $1,500,000 per bitcoin, which is unlikely because at $1,000,000 bitcoin will have replaced all other currencies.


Right now you can get an ASIC miner for $100. I'm not convinced that cost of entry will be too high for most people, though profitability might be difficult for small miners because of economies of scale. Also, the cost of power limits only the miners, not the users.

650 TWh is actually 2.8% of the world's current power production.     Also some are predicting bit coin will go up much higher than $40k, if it does clearly we could start getting into serious trouble at some point.   But there isn't any way to know what will happen with the real future value of bit coin.    

Weren't your calculations based on $.08 / KWh?   Very few get power that cheaply.    I think the most likely impact in local areas where there is a lot of bitcoin mining going on it will put pressure of electricity prices.   Of course that would mean even less power used for a $40k coin, but overall the cost would be subsidized by a local population.   I don't think that is going to be well received when it happens.   Anyway the actual picture is much more complex.      

In any case the exponential growth simply can't continue.   The growth will flatten out to a more reasonable level or else bitcoin will fail because it simply became too expensive.   I think we are already starting to see the network hash rate growth slow, but that could just be noise too.  There is also the rule of large numbers.   To grow by 30% this month  is a growth of around 4000 TH/s, when it January of 2013 there wasn't even 40 TH/s.   So it is vastly harder to achieve the same rate of growth.  

yawn PeerCoin /thread.

I agree with the "yawn PeerCoin" part.