A recent Bloomberg article stated that the Bitcoin network currently consumes about half as much energy as the Large Hadron Collider or enough to power 31,000 US homes. Even if the article is way off the mark, it is undeniable that the network's energy consumption (based on hash rate) goes up in direct proportion to the market price of a unit of the Bitcoin currency.

According to blockchain.info, the network is currently consuming 1,203.97 megawatt hours per day.

How much energy is the Bitcoin network expected to consume as the price of a bitcoin goes up from around $100 today to $1,000, $10,000, and so on?

marked as duplicate by eMansipater, David Ogren, jgm, ThePiachu, Nick ODell May 11 '13 at 16:51

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • To help answer this question, let's assume a shorter time frame and "all else being equal," i.e. no change in block creation fee (currently 25 BTC), no change in hardware efficiency, energy efficiency, and so on. In other words, let's say the market price of a coin goes up to $1,000 next month and $10,000 the month after, then calculate how much energy the network would consume given the economics. – Manish Apr 21 '13 at 16:14
  • The algorithm that blockchain.info is using had been out of date for more than a year. By far the majority of hashes today are created using ASICs which are 50X to 100X more efficient per-hash than GPUs are. – Stephen Gornick Apr 22 '13 at 0:04
  • You make a false claim that electrical consumption goes up as the price goes up. More efficient hardware causes less efficient hardware to be removed from service. So actually we are possibly already past peak electrical consumption as ASICs are forcing GPUs out of the mix. – Stephen Gornick Apr 22 '13 at 0:21
  • 1
    @StephenGornick More efficient hardware drives less efficient hardware out, but it just means more hash power is deployed (difficulty raised). The arms race continues, limited only by the cost of energy and the market price of a bitcoin. – Manish Apr 22 '13 at 2:35
up vote 4 down vote accepted

The specifics of hardware (cost efficiency and hashing rate) are not very important. The availability of the most competitive mining hardware matters, but in the long run, the cost of electricity burned will tend towards the value of BTC granted to miners. obviously it won't quite get there (they have to return their hardware costs), but barring availability issues like with today's ASIC's, it's going to get pretty close.

After all, setting up a mining operation is a no-brainer for anybody with a technical background, and hardware is a "sunk cost" - which also does nothing useful other than maintain the BTC network.

So at 25BTC per 10 minutes = 150 BTC per hour; about $18,750 worth currently. If only about half, or say $10K/h are spent on electricity costs, assuming 12¢ per kWh, that's 83MW.

A gas/oil/coal power plant gives out an output of around 1GW.

So yes if BTC goes to the ~$1k range over the next few years, it will draw about as much energy as an entire power plant produces. That's a lot of real world physical resources burned away for securing a financial transactions network with a $10B cap (at that price level).

It does get cut in half every four years though, with the halving of the block reward.

The upper boundary for energy usage can be described with a function of the profit and energy costs. Miners won't ever pay more for energy than they get from profits. The actual energy usage formula will also have to count for the hardware costs.

The formula will look something like this :

max energy usage = (profit - hardware costs) / energy cost per unit
profit = exchange rate * generated btc
generated btc = transaction fees + block creation bonus

As you can see, the energy usage depends on the following factors:

  • Energy cost per unit (kWh)
  • Hardware costs
  • Block creation bonus (currently 25, some day 0)
  • Transaction fees (currently hardly relevant, some day very important)
  • Exchange rate

And, of course :

  • How much money the miners want to keep
  • Taxes?

[edit]

$10,000 a coin times 52560 (blocks per year) times 25 (coins per block) gives $13,140,000,000 mined USD a year. Yes, that's thirteen billion USD in mining profits. If we divide this by $0.20 per kWh, then the upper boundary for energy use will be 52,560,000,000 kWh (52 TWh), roughly 2/1000 of the total energy consumption of the USA. Of course, if we assume that 90% of the money goes to hardware and not energy, the actual energy usage will be more like 5TWh. (Link to the calculation)

  • 1
    Your answer is great; but it falls just short of answering the question. All other factors remaining constant (including block reward), how much energy would the Bitcoin network consume at $10,000 a coin? I've accepted another answer, but if you wish to give it a shot I can change my vote. The formula you provide is really useful, I hope you get more upvotes. – Manish Apr 27 '13 at 23:15
  • I just updated my answer with some calculations. – Tom van der Woerdt Apr 28 '13 at 0:23
  • Great answer. It also shows that it is a real threat. The bitcoin could easily deflate to much more than $10k – Bram Nov 15 '13 at 17:51

I assert that this question is virtually irrelevant based on another assertion:

Bitcoin is more energy efficient and will always be more energy efficient than the systems it intends to wholly replace.

In both Bitcoin and the legacy currency and monetary systems, there is energy spent for some of the key tenets of a value storage and exchange system.

  • Consider the energy expended to hold balances.
    • How much electricity is used to power the banks daily? How much energy is wasted in the form of gasoline for employees' commute?
    • How much electricity is used to simply store Bitcoin data? Zero.
  • Consider the energy expended to process transactions and route money to its destination.
    • How much electricity and such is used to move money from one place to another? Sure, it's largely electronic just like Bitcoin, but let's also account for the production, movement, and destruction of physical currency.
    • How much electricity is consumed for Bitcoin mining? This increases daily, but as other answers indicate, the consumption rate will decrease if blocks aren't entering the system often enough. This is the biweekly difficulty adjustment.
  • Consider the energy expended to investigate chargebacks and fraud.
    • How much energy does it take to reverse a single credit card transaction? How many people, how many phone calls, how much electricity is wasted because of preventable fraud?
    • Bitcoin cannot be reversed, only recovered. However, the investigation could expend a lot of energy since it's like recovering physical cash.
  • Consider the energy expended to market the technology and advocate for it.
    • How much do the players in our payment system expend their own profits in order to further its system because of the business necessity to make money?
    • How much are you willing to expend to help foster Bitcoin "everywhere you want to be?"
  • 1
    This is a great argument, I'm glad someone finally put it into words. – o0'. Apr 21 '13 at 19:10
  • 2
    I like your "answer" and I have upvoted it (just because it's so awesome), but fact is it's unrelated to the question. All I'm asking is how to compute the total energy consumption given a rising market price of a bitcoin and all else being equal. – Manish Apr 21 '13 at 20:40
  • 1
    great argument, but it's also been proven wrong. BTC is now consuming as much energy as a moderately sized country. – yonil Nov 19 '17 at 23:31

There is no guarantee that transaction processing, aka "mining", will always be attractive. As such there is no guarantee that the network 'difficulty' will always go up, it fluctuates.

Therefore in a scenario, difficulty can settle at a value much lower than it is today, and/or specialized computers can become more power efficient. Bitcoin can also use a different algorithm than SHA-256, rendering sha-256 asic machines useless and no longer contributory to the network energy.

It is impossible to answer your question based on the simple premise you present.

  • Let's assume a shorter time frame. Let's say the market price of a coin goes up to $1,000 next month, $10,000 the month after, and so on. – Manish Apr 21 '13 at 16:17

In a very short amount of time essentially all mining will be on ASICs. The BFL Jalapeno just shipped. That generation of ASIC will probably represent the majority of hashing when all backorders are shipped.

That device does 5.5 Ghash/s and consumes 30W, per David Perry's first hand account of the one he received.

Today the hashing capacity is 66 Thash/s. That is 66,000 Ghash/s. If all of that was done on BFL Jalapenos there would need to be 12,000 of them. Each consumes 30W so all 12K of them draws 360kW. So in one day that electrical consumption is 8,640 kWh. At $0.15 per kWh (the average residential price of electricity in the U.S.), that totals $1,296 per-day.

The Avalon ASIC is about half as efficient (60 Ghash/s consumes 600W), so if all mining today was Avalon ASICs, then that total electrical consumption is about $2,376 per-day.

Maybe close to a third of the network is still GPU (e.g., 22Thash/s), so perhaps that would be about $35K USD per-day, but those GPUs will quickly be dropping offline as their profitability drops thanks to ASICs.

This doesn't help know what total electrical consumption a year out will be but that is because it isn't known today what the exchange rate will be a year out.

Also, today's ASICs are first generation. They could get replaced with new designs that perform 50% better even.

  • 1
    At a market price of $100 per coin, the current daily total mining reward is $360,000 @ 25 coins every 10 minutes. Total energy consumption should tend towards that amount (minus cost of hardware, operating expenses, taxes, etc.). At $0.15 per kWh you get 2,400 megawatt hours per day. – Manish Apr 22 '13 at 14:50

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.

Not the answer you're looking for? Browse other questions tagged or ask your own question.