Research shows it is hard to make money with bitcoin mining

[Jonesd]

I reverse engineered the bitcoin network mining over the last couple of years and looked at the sustainibility of it from an economic point of view. My research was just published in a peer-reviewed journal.



In short we estimate that between 2012 and 2016:

- Two billion dollar of bitcoin was mined ($2,007,950,000, based on daily price)
- The miners ran a communal deficit of approximately $558 mln at an energy cost of $0.12 per kWh. An energy price of $0.03 per kWh is required to run a small profit.
- Approximately $2.1 bln was invested in hardware. (of which $330 mln was still operational at the end of the measurement period)
- Bitcoin hash rate increased by a factor 347,000
- Electricity costs increased rapidly and amounted to almost 900 million dollars.
- The estimates of energy consumption of the network are in line with De Vries (2018) and Bitcoin Energy Consumption Index(BECI).
- Payback time for hardware was usually between 100 to 300 days, but could be as short as three days. However, his time was often too short to earn the investment back before the mining ceased to be profitable due to increased network hash rate

You can read more about it here: https://medium.com/@jonaderks/where-is-the-money-in-bitcoin-mining-64e13d2c94c0

Tell me what you think! I can imagine some findings don't match with your personal experiences.

[mocacinno]

Thanks a lot! I really loved collecting it all and see patterns emerge. Price is a big factor in bitcoin mining. During my measurement period, the price was dropping for a while, but this did not really hurt hash rate. What do you think the reson is?

I don't think there's just one single reason why the hashrate doesn't drop on a price dip. These are some of the reasons that might contribute to the phenomenon:

- a miner running an asic farm can't switch pow algo's, so he doesn't have a lot of choices to start mining on a different chain. As long as he produces more btc than his hw cost, he"s probably better off if he keeps mining... Even if he can't roi.

- some amateurs don't care about profit, they just mine to keep the diff high

- ignorance

- people that keep mining because they believe the price will rise again, but don't realise they're probably better off if they'd stop their mining and buy btc directly instead of paying more for their power than they generate in btc

[Jonesd]

I don't think there's just one single reason why the hashrate doesn't drop on a price dip. These are some of the reasons that might contribute to the phenomenon:

- a miner running an asic farm can't switch pow algo's, so he doesn't have a lot of choices to start mining on a different chain. As long as he produces more btc than his hw cost, he"s probably better off if he keeps mining... Even if he can't roi.

- some amateurs don't care about profit, they just mine to keep the diff high

- ignorance

- people that keep mining because they believe the price will rise again, but don't realise they're probably better off if they'd stop their mining and buy btc directly instead of paying more for their power than they generate in btc

I think you're close to the truth on all those points. Too bad I cannot really state such things in a peer reviewed paper.

I was also thinking that perhaps some parties have early access to new hardware and can reap the really high profits you get when hardware is much better than what's out there. What you you think of that hypothesis?

[Steamtyme]

I was also thinking that perhaps some parties have early access to new hardware and can reap the really high profits you get when hardware is much better than what's out there. What you you think of that hypothesis?

This has been speculated a lot, mostly when looking at Bitmain. It's never been legitimately proven but the tale goes they mined their hardware for months before turning around and selling it to consumers. If I remember correctly it was also visibly noted on other chains as they rolled out certain alt coin miners.

More transparently and provable on that note is the first batch orders. These were the cases you would have noted with a 3 day ROI and yes quickly became a longer timeframe.

- people that keep mining because they believe the price will rise again, but don't realise they're probably better off if they'd stop their mining and buy btc directly instead of paying more for their power than they generate in btc

I think a lot of this category are people who never attained an ROI, and are just holding out to break even in their minds. Some just like things loud and hot.

[Jonesd]

I was also thinking that perhaps some parties have early access to new hardware and can reap the really high profits you get when hardware is much better than what's out there. What you you think of that hypothesis?

This has been speculated a lot, mostly when looking at Bitmain. It's never been legitimately proven but the tale goes they mined their hardware for months before turning around and selling it to consumers. If I remember correctly it was also visibly noted on other chains as they rolled out certain alt coin miners.

More transparently and provable on that note is the first batch orders. These were the cases you would have noted with a 3 day ROI and yes quickly became a longer timeframe.

- people that keep mining because they believe the price will rise again, but don't realise they're probably better off if they'd stop their mining and buy btc directly instead of paying more for their power than they generate in btc

I think a lot of this category are people who never attained an ROI, and are just holding out to break even in their minds. Some just like things loud and hot.

The hardware companies first using their machines would make a lot of sense. As they have to test the machines anyway, they should have early access anyway. You would expect something like this to be buried in pool archives somewhere (as they show the block winners, but likely anonymous accounts).

The first batch orders make sense, indeed.

On the longer term, you really wonder how much more centralized mining will become. It would make sense the hardware companies becoming more powerful. Just look at something like Hut8: https://hut8mining.com/investors

Think bitcoin should switch to something ASIC resistant?

[Steamtyme]

I don't think switching to an asic resistant algorithm would make a difference.

Some of the coins that where supposed to have this feature have already shown that once they are worth it an asic is developed. That is one benefit to the fact that there are stubborn miners out there or even enough larger miners of differing views to keep the hash fairly spread out.

I'm glad to see more hardware manufacturers have entered the space, as they will obviously compete against one another. This will hopefully help maintain a somewhat decentralized network. In the long run though I unfortunately see things becoming more of a "board" mentality as rising difficulty will phase out a large portion of the population from being able to economically run equipment. This will widdle things down to a few outliers but essentially a handful of extremely large scale miners, who in the end will likely be made up of the manufacturers.

[buwaytress]

Nice article, OP. Be curious to go elbow deep into some of your numbers together with some miners. You admit yourself that there is no factual data for your estimation of installation costs. There are surprisingly a number of independent people in the Netherlands itself (go Vrije!) still starting out on their own rigs and I personally suspect that there are still so many ways to shave costs perhaps including.

1. Cost of the rigs: We assume people buy new ones off the shelf, but it is possible many more serious miners are also able to build some of their own parts or access at even cheaper costs, or at least, below market.
2. Sale of Bitcoin: We assume people sell at market price, but it is possible also they regularly trade OTC or even between the spreads you are using, that's an extra 2-3% at least on income.
3. Energy use: We assume people are paying honest market prices, but we know, for example in China, majority miners in those periods were able to cut direct deals with energy companies for very good rates, or it could be they are buying at cheaper rates with various subsidies.

Mining was never for the faint-hearted, but I think we know a bit less than what they're letting on, those profitable ones. But of course, all this buying and feeding the manufacturers just allows them to make even more money, not necessarily anymore from mining itself.

[fanatic26_]

There are certain times where it makes sense to get into mining, and other times when it doesnt. When the prices started going crazy last year there was a window to start something and recoup your costs before everything went back to normal. In lean times like these, mining only makes sense if you already have the infrastructure in place to support it.

P.S. To the comments about making bitcoin anti asic. This shouldnt even be discussed as a remote possibility. This isnt some random shitcoin that nobody cares about, this is the largest processing network in the world. There are hundreds of billions of dollars in infrastructure and products on the market to support it, you cant just flip a switch and turn it all off. Bitcoin as its known now would die, plain and simple.

[ccgllc]

I don't think switching to an asic resistant algorithm would make a difference.

Some of the coins that where supposed to have this feature have already shown that once they are worth it an asic is developed. That is one benefit to the fact that there are stubborn miners out there or even enough larger miners of differing views to keep the hash fairly spread out.

I'm glad to see more hardware manufacturers have entered the space, as they will obviously compete against one another. This will hopefully help maintain a somewhat decentralized network. In the long run though I unfortunately see things becoming more of a "board" mentality as rising difficulty will phase out a large portion of the population from being able to economically run equipment. This will widdle things down to a few outliers but essentially a handful of extremely large scale miners, who in the end will likely be made up of the manufacturers.

There are some little known coins, like IMACredit (www.imacredit.org) that remain ASIC resistant due to large memory footprints.  Alas, they were too "different" back in 2014 to be accepted by the major exchanges.  People back then couldn't wrap their heads around a coin that hashed in the 100s of hashs a second instead of in the megahash realm.  You know... like zcash SOLs do today?  Alas, they were written by smaller groups of developers and never caught on - although a few diehards still CPU mine them for the fun of it.

[squall1066]

its all down to your personal philosophy, personally, I mine for coins now for the price in the future, I DO NOT run more units than I cant afford to pay for in fiat, if you run units to pay for them selves there WILL be periods of "loss" as you will not mine enough to cover running costs.

this is true providing you don't have free power (or incredibly low) right now I don't understand how 3/4 of the world in mining, the price is marginally above reward over running cost, so there is next to no chance of ROI your unit?

anyone know the story of the gold rush? more money was made by the people selling shovels than people made in collecting gold dust. my advice, if you cant run a unit, just buy what you can every week/month and sit on it :-)

[mikeywith]

The hardware companies first using their machines would make a lot of sense. As they have to test the machines anyway, they should have early access anyway. You would expect something like this to be buried in pool archives somewhere (as they show the block winners, but likely anonymous accounts).

[...]

I disagree with this theory of hardware companies use their machines before selling them. unless you mean " using their own , keeping their own" and after wards sell " other machines of same specs" that COULD make sense.  but there is noway they could mine with an ASIC and then pack it, and sell it as a brand new.

if you have ordered any ASICs from bitmain "for example" you could tell by

 1- how clean they are: I randomly disassemble my gears and they all look 100% dust-free.  the cost and time needed to bring it back to this state after even as low as 1 day of mining is amazingly high. try it out, it just will not work. a brand new device is a brand new device.  

 2- the smell of the "never used"electronics. if you have enough experience dealing with electronics you can really SMELL and FEEL something that has never been powered up, not even for testing.

of course there could be some exceptions of some ASICs they used/tested and then sold as a brand-new but this is minimal and most people who support this claim against hardware companies are people who lost money investing in something they do not understand and will do/say anything to make themselves feel better. not defending any ASIC manufacturer just stating facts    

aside from that I agree with your concepts. very nice paper and good job.

oh i almost forgot ,what do you mean by

"More efficient consensus mechanism like Proof-of-Stake to counter decentralization that results from high upfront investments and risk."
is required for bitcoin mining to be a sustainable business (in an economic sense) ?

[Jonesd]

[...]

I do not mean they resell those machines, but they could use first shipments for theirselves when the competition is still low and than sell later shipments to the general public. Reselling used stuff would be dumb, indeed

Thanks for the feedback

I mean with that last statement that the energy costs and hardware costs are much lower with PoS and thus me efficient. Security-wise it's a whole other debate of course.

its all down to your personal philosophy, personally, I mine for coins now for the price in the future, I DO NOT run more units than I cant afford to pay for in fiat, if you run units to pay for them selves there WILL be periods of "loss" as you will not mine enough to cover running costs.

this is true providing you don't have free power (or incredibly low) right now I don't understand how 3/4 of the world in mining, the price is marginally above reward over running cost, so there is next to no chance of ROI your unit?

anyone know the story of the gold rush? more money was made by the people selling shovels than people made in collecting gold dust. my advice, if you cant run a unit, just buy what you can every week/month and sit on it :-)

The chance of running a profit in the long run is pretty slim, but if you happen to be there in periods when price rises very fast and everybody is paying top dollar for the bitcoins, you can earn the machine back in days. The thing is that you could have also just bought some bitcoin and have perhaps an even higher ROI over that period and much less of a hassle.

[philipma1957]

... The chance of running a profit in the long run is pretty slim, but if you happen to be there in periods when price rises very fast and everybody is paying top dollar for the bitcoins, you can earn the machine back in days. The thing is that you could have also just bought some bitcoin and have perhaps an even higher ROI over that period and much less of a hassle.

My chances of turning a profit are close to certain.  I would say 95% is a conservative number. For me.

most miners don’t do great they lose.  Mining for the most part is cheap power = king.

I have one solar array deal.  And I am going into a second solar array deal.

The first one does 15kwatt 24/7/365 on average. The second one will do 40kwatt 24/7/365

Total of 55kwatt per hour about 600th for btc and 6.5 gh for ltc.

Not a lot of profit but profit.  Which should not stop.

I am not the only miner with good power. I know lots of people with 2-4 cent power  maybe 300megawatts

Which is a lot of btc hash.

They will earn.  Will A guy in his home with 2 x 30amp 240 volt circuits make money not so easy or certain for him.

[Jonesd]

My chances of turning a profit are close to certain.  I would say 95% is a conservative number. For me.

...

They will earn.  Will A guy in his home with 2 x 30amp 240 volt circuits make money not so easy or certain for him.

Very cheap electricity certainly makes a difference. What about the payback time on your machines? How long in your experience before getting your initial investment back?

[philipma1957]

Very cheap electricity certainly makes a difference. What about the payback time on your machines? How long in your experience before getting your initial investment back?

As always it varies.

Right now I buy older cheap used gear such as the s9  set with new firmware to middle speed.

1280 watts does 13500 gh

I purchased 30 new apw3++ psu units for 45 each due to a hotter going bankrupt
I purchased 19 s9s for 200 or less

for sake of argument say 250 for a unit and a psu

https://bitcoinwisdom.com/bitcoin/difficulty  plug in 13.5 th and you get 61.65 a month  or 308 in 5 months

that means those s9's will be paid off pretty fast.

Once the power has been setup it is under 1 penny a kwatt to maintain

many people have gone out and grabbed margin s9's and cheap apw3++ psus

they may  go the route of over clocking them  since they may have paid under 200 for a unit and a psu.

plug in 17th

https://bitcoinwisdom.com/bitcoin/difficulty

it is 77.63 a month x 5 = 390 in five months.  pre power.   but a 2 cent or 3 cent guy is fine with this.

this is why the network is crawling sideways as you are down to a lot of low power guys using older s9's with various clocks  as low as 9thand as high as 17th to suit their setups.

If they are like me they slowly upgrade to s15 or t15 or m10 while earning with cheap used s9's

anyone under 4.5 cents can do this.

https://bitcoinwisdom.com/bitcoin/calculator

see calculator above.

results are from using sideways diff of 1 %  power of 4.5 cents
a 200 usd s9 and psu

Date   Difficulty   Revenue   Profit   Return

2019

2-2 – 2-28 (27 days)   5814 G   17.2   17.03   -183
3-1 – 3-31 (57 days)   5931 G   18.76   18.57   -164.4
4-1 – 4-30 (87 days)   6050 G   17.23   17.06   -147.3
5-1 – 5-31 (118 days)   6172 G   16.85   16.68   -130.7
6-1 – 6-30 (148 days)   6296 G   15.38   15.23   -115.4
7-1 – 7-31 (179 days)   6487 G   14.95   14.8   -100.6
8-1 – 8-31 (210 days)   6617 G   14.04   13.9   -86.72
9-1 – 9-30 (240 days)   6750 G   12.74   12.61   -74.11
10-1 – 10-31 (271 days)   6886 G   12.3   12.18   -61.93
11-1 – 11-30 (302 days)   7024 G   11.1   10.99   -50.95
12-1 – 12-31 (333 days)   7237 G   10.6   10.49   -40.45

2020

1-1 – 1-31 (364 days)   7383 G   9.785   9.687   -30.77
2-1 – 2-29 (393 days)   7531 G   8.427   8.343   -22.42
3-1 – 3-31 (423 days)   7682 G   8.239   8.157   -14.27
4-1 – 4-30 (453 days)   7837 G   7.251   7.179   -7.088
5-1 – 5-31 (484 days)   7994 G   6.733   6.666   -0.4217
6-1 – 6-30 (514 days)   8237 G   5.812   5.754   5.332

now that is at the low power setting of about 80 watts a th

and that is if you buy the s9 now.

if you have it already the gear project to earn 205 use after 4.5 cent power cost over next 514 days.

if you have 100 units paid off  you would turn a 100 x 85 = 8500 profit   from now to june 30  so you could swap in new gear once a month  an s15 or t15 or m10

which range from 50 to 70 watts a th

[Steamtyme]

Will A guy in his home with 2 x 30amp 240 volt circuits make money not so easy or certain for him.

Sounds like a guy I know.

In my case to make that scenario work, my miners are heaters. This obviously only applies to Winter months but it's a straight trade off versus my electric furnace.

The best way to squeeze a profit out in this market is to get creative. I don't like the idea of it but I could see a lot of home miners becoming seasonal, if the conditions are right.

My situation currently has these units supplementing the heat in the home.

4 * 741 = 31 TH/S - these would be otherwise obsolete
2 * m10 = 48 TH/S - 1 of these is in low power mode to manage in house noise levels.

waiting to find a spot is my 841 that is in the garage until further notice

[sidehack]

Sorry to interject, but every time I see the subject line of this thread pop up I can't help but chuckle and think "Research shows peanut butter is sticky" or "Research shows fire can be quite hot".

[philipma1957]

Sorry to interject, but every time I see the subject line of this thread pop up I can't help but chuckle and think "Research shows peanut butter is sticky" or "Research shows fire can be quite hot".

Water is quite wet except when it is hard and frozen.

[Artemis3]

I do not mean they resell those machines, but they could use first shipments for theirselves when the competition is still low and than sell later shipments to the general public. Reselling used stuff would be dumb, indeed

Oh but they do resell them, after becoming obsolete... Or, they start a "cloud mining business" where they sent them to.

[Jonesd]

Sorry to interject, but every time I see the subject line of this thread pop up I can't help but chuckle and think "Research shows peanut butter is sticky" or "Research shows fire can be quite hot".

Well, I can tell you that it is quite hard to get research peer reviewed and published in an academic journal. One of my goals a couple of years ago was to get the academic world to take crypto seriously and get something published.

Looking at the investments people make without really looking into the risk, makes it clear that a lot of people assume it is easy money. I don't like to assume too much and study the data.