Electricity consumption of Bitcoin: a market-based and technical analysis

[mrb]

I thought you guys might be interested in a technical analysis of the electricity consumption of Bitcoin miners:
http://blog.zorinaq.com/bitcoin-electricity-consumption/

[BurtW]

I really appreciate your careful calculations and the following summary:

Lower bound   Best guess              Upper bound
325 MW          470-540 MW           774 MW
2.85 TWh/yr    4.12-4.73 TWh/yr   6.78 TWh/yr
0.100 J/GH      0.145-0.166 J/GH   0.238 J/GH
$142M/yr        $206-237M/yr        $339M/yr

For some perspective, this article:

https://en.wikipedia.org/wiki/World_energy_consumption

Puts the 2013 worldwide power consumption at about 12.3 TW.  So using the numbers above we get:

Lower bound   Best guess               Upper bound
0.00264%      0.00382-0.00439%   0.00629%

of worldwide power consumption.

[mrb]

That's right BurtW. In the past I made this comparison in another post (http://blog.zorinaq.com/bitcoin-mining-is-not-wasteful/). I added these percentages to this post, updated with IEA 2014 stats.

[amaclin]

Can we calculate from these digits and plot a chart how much electricity bitcoin network consumes
a) daily
b) per one transaction confirmed in block
?

Oups, sorry... The answer can be calculated from the digits on your page

[d5000]

Good work! I've long searched for a well-researched study of this topic, and all I found were the flawed "price-based" estimations.

Now I've got a question: Is mining farms' electricity consumption for cooling included in the study or is the study based on the raw electricity consumption of the miner hardware? Cooling is mentioned in the article but I couldn't say for sure if the cooling costs are included in the J/GH calculations that are the base for the study.

It surprised me to not see cooling costs as a separate "cost item", as there should be different configuration options for cooling in mining farms and also different challenges depending on the climate of the location. But I'm no expert here as I never tried to mine (at least not BTC ).

[mrb]

Indeed, I also wrote a critic of one of these price-based estimations: http://blog.zorinaq.com/serious-faults-in-beci/

First of all, note that the majority of mining farms operate at very low overheads—PUE 1.05 or less. I start my calculations by taking into account only system-level cooling overhead (chassis fans, Bitfury liquid cooling tech, etc), and not data center-level cooling overhead, because my power and efficiency figures are measured "at the wall".

For my "best guess" figures, in theory I should take into account the PUE by adding 5% to my figures. But the range is pretty large (470-540 MW), not meant to be ±5% accurate, so I don't think it's useful to add 5%.

For my lower and upper bound, I don't need to take into account the PUE. The lower bound, by nature, needs to assume the overhead is zero. And the upper bound is such a worst-case & unrealistic estimate (by assuming miners always deploy the least efficient hardware of their time) that it must surely already overestimate power by at least 5%, which is equivalent to taking the PUE into account.

[d5000]

Thanks for clarification! Your methodology seems to be totally legit if the PUE of mining data centers really is only 1.05 - because, as you say, the difference would be less than the variance.

The only thing that could be criticised is that maybe the figures delivered by the mining equipment manufacturers could under-estimate the power usage "in the wild" (like we know it from fuel consumption by cars). But I think that should not be that relevant because if they would underestimate it too much miners would have already challenged their estimations (or even engaged them in legal problems).

Still I consider this work the best on the topic.

[mrb]

Thanks d5000! I put a lot of effort into gathering the data, making sure it is correct, and writing this post. I am glad you appreciate the research I put into it.

In general I took the power efficiency numbers as reported by end-users (when available), which tend to slightly over-estimate manufacturer figures. Eg. the Avalon 6 was advertised at 0.28 J/GH, but I use 0.29 J/GH as reported by users with their power meters.

[mrb]

I published my article in Bitcoin Magazine:
https://bitcoinmagazine.com/articles/op-ed-bitcoin-miners-consume-reasonable-amount-energy-and-its-all-worth-it/

[spin]

Refreshing to have some fact based work to look at make it refreshing.

I've used your numbers to redo your calculations slightly differently:

Firstly I assume we invest BTC to generate a BTC profit .

This is as opposed to measuring a USD profit.  The idea being to make the returns a little less sensitive to the exchange rate increasing.

I assumed the $418 spent on the miner was converted to 1.33 BTC (using the 315.17 USD/BTC rate you have in the CSV).
I convert the electricity cost to BTC using the daily rate.  
Total BTC generated = 5.07
Total electricity is BTC1.53
Net income = BTC3.53

So 1.33 turns into 3.53 BTC for a gain of 2.66.

This is still sensitive to price of BTC.

Assume flat prices at 315 USD/BTC

To get an idea what the sensitivity of this income was I set it to a flat 315 exchnage rate. This assume a flat exchange rate throughout the period in question.
1.33 then turns into 3.24 giving a 2.45 gain

Assume flat prices at 1000 USD/BTC
0.42 BTC then turns into 4.5 BTC for a gain of 10.76 times.

[mrb]

Interesting, nice calculations.

Ultimately everything is always linked to the BTC price. If it doesn't gain value, fewer people are enticed to mine, so the difficulty doesn't increase as much, leading to more BTC-denominated profits.

[mrb]

I have updated my models (CSV files) showing how much was mined in dollars as well as in BTC by the Antminer S5 since its release date, and I also added data for the S7 and S9:

http://blog.zorinaq.com/assets/income-antminer-s5.csv
http://blog.zorinaq.com/assets/income-antminer-s7.csv
http://blog.zorinaq.com/assets/income-antminer-s9.csv

Across its lifetime (which ended on 8 Oct 2016 as its first unprofitable day), an S5 mined 4.93 BTC ($1021 if selling the BTC on a daily basis).

Across its lifetime (still profitable and running as of 15 May 2017), an S7 mined 6.96 BTC ($2481 if selling the BTC on a daily basis).

Across its lifetime (still profitable and running as of 15 May 2017), an S9 mined 4.56 BTC ($2974 if selling the BTC on a daily basis).

(Note that these calculations do not take into account extra income from tx fees, as until recently most pools did not pay out fees to miners.)