Bitcoin Forum

Most efficient bitcoin miner - Extolabs EX1/ Wolfblood-XE  - 1 GH/s for 0.19 watt

                                                                                             5263 GH/ for 1 kilowatt

Cheapest Electricity in USA/Europe/Russia/China (avg)     - 0.08 USD/kWh


Estimated time to generate 1 BTC                                    - 10.36 days (with the current difficulty)

Your math is wrong. It takes 259 days on average to mine a whole block of 25 BTC, so energy cost per BTC for this miner is only 19.89 USD @ 0.08 USD/kWH. We have some ways to go before those become unprofitable. ;)

Eh...I pay $0.08/KW in the US and I live in a highly populated area that is NOT cheap. I have a buddy in the Mid west that pays $0.04/KWH (has agriculture line) and I've heard rumors that some big centers in china can get $0.02/KWh.

I'm paying $0.26 USD/kWH here where I live, I feel scammed. :(

At the current times 25x profitability in mining, that's huge

Yes, nethash has to rise a lot to create again a reasonable price equilibrium. Mining is a gold laying chicken for the professionals.

you see bitcoin is going to nowhere but down,bitcoin is too much over-price !

That would be good if miner was free :D

You have to take cost of the miner and how much it could produce in certain time frame, add electricity costs to that, and then you'll get mining profitability.

I seriously can't understand why anyone beside some huge entity that could get a good mining hardware deal or produce it's own, would buy a miner now.

+1 for seleme's point.

Let's do the calculation right:

Hardware cost + total energy cost over entire life time of (energy profitable) mining of hardware, divided by coins mined during that time = price per coin. Note that this calculation has to include an estimate of difficulty growth, to get a realistic value for 'life time of energy profitable mining'.

Two further remarks:

- I did my own calculations about 2 months ago and got a value around $470, assuming low consumer energy costs and low consumer hardware costs. However, there was and is a big unknown element in calculations like this: We don't know the factor by which large, professional mining farms can get hardware and energy at a discount.

- For energy, about $0.05/kWh seems a low, but attainable estimate.

oda.krell:
Interesting bit re. rationality of miners/tx fees (from http://www.coindesk.com/faster-bitcoin-network-keep-fees-low )


Another tangent re. "Hardware cost + total energy cost over entire life time of (energy profitable) mining of hardware, divided by coins mined during that time = price per coin":

That is the point at which a rational player would start mining.
Once the mining gear is bought, he will continue mining until Hardware cost + energy_cost_to_mine_coin > price_per_coin.

Yes, but impact of energy cost only  mining decreases over time, as the hashrate share of total hashrate declines . Exponentially, in fact.

There's really no way around it: cost per coin is based on total coins produced and total cost during production. The latter includes hardware cost.

Edit : to clarify. Because of the hashrate increase, for any given fixed hashrate miner, a finite amount of coins is generated by that miner. (or, more accurately  I suppose : that number of coins is approached asymptotically

Ah, speculation, the brain damaged section of bitcointalk.

0.19W/GH sounds pretty nice. Especially considering the currently best models require ~0.5W/GH. Just looking in the mining sections provides the following information about Extolabs (https://bitcointalk.org/index.php?topic=583857.0). The web page (http://www.extolabs.com/) also doesnt exist anymore.

Calculating current costs using the efficiency of a scam product. Very good. And, as others have mentioned, you need to checkthe "whole package" to calculate profitability (up front cost, electricity cost over whole timeframe, maximum btc earned).

oda.krell:

What I'm trying to say is there are scenarios (not just hypothetical) in which a rational miner is forced to continue mining even though overall, the operation is unprofitable.
Having a hard time putting it into words, maybe an example might be clearer:

1.  I spend $1k on hardware.  At current difficulty and [predicted] difficulty increase rate, I calculate that ((hardware cost) + (hosting cost)) < (price of coins mined over lifetime).  Profit.

2.  Other rational miners do same, difficulty skyrockets. <=="turn on" calculations, using your formula, all made sense.  Rational.

3.  Now my original calculation is no longer true.  I will lose money, but wat do?  My options are as follows: <=="turn off" condition
   3a.  Shut down as soon as my original calculation, (1), is no longer valid, losing my original 1k investment.
   3b.  Keep mining while (hosting cost to mine one coin) remains < (price I could get for that coin).  <==obviously the rational choice, my loss will not be total, but I won't break even either.
   3c. ???

Is this any clearer? (it's morning :()

Yes :)

But you're calculating now (average)  profit of miners, while I talked about _cost_ per coin.

I never said price (market price, that is) can never fall below production cost, which is what you have in mind maybe. That said, I believe production cost is a price attractor, and we won't fall below it unless extremely strong momentum brings us there. If price would fall below production cost for a longer period, we would certainly see the effect of that in hashrate / difficulty, but we don't (yet).

No matter how cheap one's electricity is, if the hardware cost is taken into account, one always loses money by mining!!!! Sorry!

Where I live the electricity costs are about 0,30 USD/kWh, I wish I would live in China and own a big bitcoin farm...

Probably true for the majority of amateur miners.

Almost certainly not true for the professional mining farms.

I used many calculators on the net and even used $0.02/kwh. If difficulty keeps increasing at the current rate (20% per adjustment), nobody is profitable one year from their purchase.

My point is the price feedback loop is laggy--it has a time delay.
Introduce a time delay into any negative feedback loop (think typical op amp circuit, or a governor on a steam engine) and what you get is an oscillator (hunting or surging for mechanical stuff).  Depending on initial conditions, the oscillation will be sustained, damped, or terminal (things go poof!)

*The time constant was really out of hand when people were preordering ASICs and getting them a year later.  

And coffee is not working for me today :(

P.S:  Do you see what I mean by sloppy price feedback?

You are masterizer of the english language!!

I quickly ran the numbers for supposedly the most efficient advertised hardware available (Spondoolies SP20 and the SP31).

Even if they are able to produce these for $100 per unit, by the time you take into consideration $485/BTC, pay for shipping and a power supply they are not profitable to run at the lowest energy costs in the country ($0.05/KW hr)  Not for one week....not for 6 months. It doesn't help to scale up, my calculation does not take into consideration commercial datacenter costs...so 1000 of these doesn't help, it makes the losses much worse.

I seriously doubt that ANY company producing hardware is able to produce 1.7 TH for $100.
(ASICMiner for example has a chip only cost is higher than that.)

There is no mining equipment produced today that can make profit, even at manufacturers prices.

You can trick yourself into believing there is a meager profit to be squeezed out if you manipulate the online mining calcultors...you are only fooling yourself.

(Not sure how the OP came to calculate what he/she did but they are clueless.)

^^^ This!!

Yes, I left the delay part out because I wasn't sure if we had the same idea about this mechanism. So, agreed on your point. However (and correct me if I'm wrong), eventually, difficulty will need to decline (or at least rate of growth will) if mining is unprofitable - together with assumption of course that miners are, on average, ecomomically rational.

What I mean is: up to now, it is possible that we don't see the effects of unprofitable mining in difficulty because of the delay in the feedback loop, but the longer there is no effect on the difficulty, the less likely it becomes that mining is unprofitable (unless we assume a growing delay, which I don't find plausible.)

Correct?

The rate of difficulty growth (and the rate of growth of mining power as well) has declined. It was +30% every difficulty correction. Now it's more like +20% or less.

I don't know how you got your numbers. Let's base the calculation on Antminer's S3. Specs used for calculations as taken from their site, see here (https://i.imgur.com/pFDjTFC.png).

Let's assume difficulty increase of 15% per adjustment, and as energy cost: $0.05 per kWh.

If mining starts now, then after 198 days, the unit will stop mining at a profit based on ongoing energy costs (assuming $500 per BTC market price), and will have generated a total of 0.64 BTC during its (profitable) lifetime.


Now, given the raw numbers so far, we get:

(((0.58*480)+(198*0.43)))/0.64 ~= $568 production cost per 1 BTC.

Which is indeed higher than current market price, and suggests mining isn't profitable anymore.


However, keep in mind the point I've made several times by now: amateur mining is probably unprofitable. It's the professional, large scale mining that is interesting, under the assumption of reduced hardware and energy costs.

So, I'll make one more assumption: Let's say Antminer has a 25% profit margin on their hardware going for consumer prices, that a large miner (maybe Antminer running a farm themselves) is able to deduct.

You can reject this as pure conjecture, please feel free to do so. I however consider this to be a rather conservative estimate. In this case, we get:


(((0.58*0.75*480)+(198*0.43)))/0.64 ~= $459 per 1 BTC.


Now, to get back to the point I've been trying to make over and over again:

There are several unknown variables that apply to large scale mining, such as: cost of operation (manpower, rent, cooling, setup, organizational overhead), cost of acquisition of hardware, and finally: energy cost. For that reason, I don't make any claim that the above value (of $459) is "the correct" production cost per coin of a large mining operation.

My goal in the above calculation is simply to point out, that, given the relatively conservative assumption that a large mining farm can acquire S3s for 75% of the consumer price, and make use of electricity for $0.05 per kWh, the production cost (taking into account hardware cost and energy only) is below current market, i.o.w. mining would be profitable for such an operation.

Here i made a bad chart with paint( ::)) about it:

Nice job but it only reflects the cost in electricity. Why don't you redo it adding the cost of the miner! We would have the whole picture then!

Here is a chart with the addition of the miner costs:


In the chart you see the bitcoins that the miner creates over its profitable lifetime, if you would start from the beginning of the launch. The cost of the miner is calculated by $cost/485(<-current bitcoin price).

0.2W/GH/s seems to be quite low, that's a given. What needs to be considered as well is the initial price of the miner itself. Those miners don't grow on trees. But it shows how manufacturers can make quite a decent return if they're able to churn those miners out at a low price!

A 453GH/s miner like antminer s3+ will produce 0.79983BTC over a year if difficulty adjusts by +15% every time. It costs 0.58BTC as per https://bitmaintech.com/productDetail.htm?pid=00020140813125800214NBU85kWt0672


It needs 355 watts of power which will cost $155.49 at $0.05 per kwh.
So, 0.79983BTC-0.58BTC leaves us with 0.21983BTC at $500 per BTC, it's $109.915, which is less than the electricity cost. Unprofitable!!

If we had a 25% discount from the hardware cost, we will make ~$28 by the end of the year.

All of this assumes no pool fees of any kind.

Are the Extolabs actually legit? They first came up in May if I remember correctly. The prices back then are just too high at the current levels of difficulty, though. You need to buy the miner first!

Nothing to argue there against.
It doesnt tell us anything about how much an Antminer actually costs to be built.
When announced (http://www.coindesk.com/bitmain-releases-energy-efficient-478ghs-antminer-s3/) the S3 used to cost 0.75BTC (and BTC was like $600 back then), so basicly the price has dropped from $450 in June/July to $280 now. In other words the price has already dropped nearly 40%.
Its fair to assume that the S3 definately costs less than $450 to build, otherwise one wouldnt consider building such a device. Is it still profitable to build at $280? Or is that already a clearance sale?
Probably still profitable, but unless some insider can tell us how much those companies pay for the wafers and how high the chip yield is we cant know for sure.

My guess would be the manufacturers operating their own pools are still turning a profit, but i doubt their margins are still healthy given risks involved.

You're re-doing my calculation from above, but you let the hypothetical S3 run past its profitable range wrt energy.

My cutoff was ~200 days, in which case the total cost to produce 1 BTC with a 25% off S3 is $459 (this is the price per 1 BTC - so you will need more than 1 S3 running to get 1 full BTC after 199 days, to be clear).

Tell me again how this is unprofitable?

Now we're talking. I don't think the margins are as good as they used to be for the big operators (that part is obvious, given lower price and higher difficulty), and are maybe even getting close to a net zero.

I have the suspicion (that I can't prove) that we are currently pretty close to the cost of a coin mined by a large farm. I don't think it is coincidence that the mid to high 400s are revisited so often and persistently. But it's nothing more than a suspicion, I admit. Maybe post hoc reasoning to explain why price keeps coming back to where we are now.

I'm saying that with 25% off the hardware cost, there is a tiny profit to be made. Pool fees excluded of course!!!!!

I agreed with that (the reduced margins) in my reply to cow-herder. Just as an aside, a sufficiently large farm won't need a pool, so not necessarily needs to pay a pool fee.

Anyway, in reality, I'm sure the large operators' calculations are a lot more complex (since they presumably replace outdated hardware on the go, and need to take into account how much of their coins they can sell without depressing price too much)

But my point remains that, assuming you will be able to sell over the next 200 days for a price of around $500 per coin, investing $459 doesn't seem such a bad deal to me.  This would be a ~9% profit, that's hardly "tiny". (EDIT: at $500 per coin, that is. Profit would have been a lot higher at previous prices. Point is that, even now, there'd be 9% profit)

Oda...I respect your calculation...but it is simplistic and lacking several important variables as you mentioned yourself.

There is a transition from amateur miner to commercial miner that occurs around 2400w of draw.
No one is pulling that much current from a single household circuit unless you have modified a dryer plug or something illegal.
If you burn your house down mining 0.1 btc per month...well good for you, great job.

Amateur miners may be able to eek out a few dollars of profit in 6 months of mining with the most efficient mining equipment on the market....I'm still not sure of that...most of these "mining rigs" require the amateur users to buy the power supply separately, and you have to pay duty and import taxes. This can double the hardware investment price before it ever reaches the wall socket. (For my KNC unit last year the power supply was $192, and the shipping, duty and taxes was almost $450.)

However, the biggest variable cost though occurs once you go beyond 2400w of power requirements....now the miner's have to move the mining units out of their mom's basement and go to a commercial installation.

Air conditioning alone almost doubles the electrical cost, not to mention...rent, electricians to install and sign off on wiring panels, networking demands that go beyond a few 8 port routers tied together, internet bandwidth charges that are more than your usual household data plans....the costs are exponentially higher than someone running up to 5TH in their basement.

The upside? If they can get the hardware or manufacture the hardware at cost (<20% of retail), they can reduce some of the depreciation costs.



I still put forth the thesis that there currently is NO mining hardware that is profitable being produced today, not at any scale.
If anything, the large commercial mines are running at bigger losses than the little guys like me.

Once we admit that, we can begin to think about what problems and possibly opportunities that brings.
The reason the difficulty keeps increasing is not because it is profitable, it is because of the lengthy feedback loop and the fact that we are engaged in a silent war of attrition.

I believe the feedback loop is 2-3 months...the time it takes to get new hardware off the drawing board and into the mines.

Another point.

With all mining you need to be the first to have the miner to get good profits.
This means you need to preorder possible scam products and hope they produce it and deliver it on time for it to be worth while.

In essence there is also risk in mining upping the price aswell.

Hardware depreciation cost is a large factor on why so many amateurs miners lost a great deal of money on mining.

If they use the same amount of money to buy coin directly, risk is about same but they get more benefit from rising value of btc.

is this thread for idiots?

OP must have been extremely retarded. Why not use Unicorn OP? I have  one that hash at 5PH/s and only use 5kw. PM me dumbass.

1 PH/s per kW . Is it real ? sources ?

I just wanted to know, what was the mining cost of cheapest bitcoin, mined at industrial scale. According to oda its 459 USD/BTC.

Hm. Not too happy about the way my post was understood then. I probably should have phrased it more clearly...

Like I said in my calculation post, I don't want to claim this is the correct lowest production cost per coin.

What I intend it to be was a "model" calculation -  to show how, with a relatively conservative estimates on the discounts a large mining operator could get, mining could still be profitable assuming those rebates, even though mining for "amateurs" (who don't get those rebates) is probably not profitable at the moment.

As an example, say the large miners only can eek out a 10% discount for the hardware (vs. the 25% I assumed), and their overhead cost (facilities to run miners, cooling, etc) is higher than I think, in which case even the large miners don't mine profitable at the moment. That is the point CMMPro is making (correct me please if I'm wrong here, CMMPro).

My point is: I believe it is at the very least possible that large miners get substantial enough rebates on hardware, and are efficient enough in operating, that they can still turn a profit. Especially considering that large investments continue being made in mining farms.


To condense the point I tried to make even further...

Group 1: "Bitcoin mining is definitely NOT profitable at current prices!"

Group 2: "Bitcoin mining IS definitely profitable at current prices."

My position: "I can't say whether it's profitable for sure, because I don't know the discount factor of large farms, but large scale mining might be profitable, even at current prices, if those rebates are high enough."

My experience in designing niche market consumer electronics (for racing applications) is that the production costs for small runs is about 20% of retail price. So for a $500 transmission controller for a rally car, my costs would be in the neighbourhood of $100-120 per unit. (Parts, pc boards, assembly labour...and mostly done in China.)

I believe it would be difficult for them to get much more of a discount, since my productions numbers were maybe half what theirs are, and my parts count was lower as well but generally I think we can assume somewhere between 10-20% of retail price is what one of the mining hardware manufacturers are paying for their hardware.

So, we can estimate that for a 5.5TH miner that retails for $4000....they are paying minimum of about $400 per unit cost.
Add 3kW of power supply and maybe you are at I dunno maybe $700(?)

(Edit: Whoops was I wrong on the power supply cost...they can be $300 used and $2000 new.)
Good luck finding 500-1000 used 3KW power supplies too!

Add rent and property taxes on the building...difficult to determine.

Add infrastructure....large scale power transformers, electricians, lan cabling, lan distribution, internet bandwidth

Add power costs...$0.05/kw(?) Not sure when you are buying in megawatts per month.

^Double that for air conditioning and fans to cool the place.

Add a salary 24/7 to maintain it when units go down...

Budget for maybe 5% spares of mining hardware and power supplies when units break down.


I can't see it making money unless some of these costs are eliminated.

But, then again: what is the "real" price of a miner? All of them get 50% cheaper just 2-3 months later. So, is the real retail price 4000$? Or 2000$? Or even 500$? We just don't know without insider informations.

It is a really scary prospect. They day mining cost becomes unprofitable (soon), there will be a lot of dead mining companies thus an huge flash crash, thus killing even more miners. Also no one will lend money to this companies during theses times.

What if you get free electricity and pay nothing for it? There are some places where stealing power is no big deal. Like where I live.

Just a quick remark re: operating costs.

What do you think the wage is these guys get paid?

https://i.imgur.com/F6dITai.jpg

living in here when they're on duty

https://i.imgur.com/aV6HARG.jpg

from this article: http://www.thecoinsman.com/2014/08/bitcoin/inside-chinese-bitcoin-mine/


I get the feeling to accurately estimate the lowest cost to produce coins, we need to get prices found in the US/Europe out of our head, and instead look at the Chinese market instead.

Nah.  Many mining companies have been running in the red (if you include the cost of gear) for quite a while.  Look at stuff like PETA, SCRYPT, HMF, Labrat.  These are all for-profit "professional" mines.  They will continue operating even when electrical costs surpass the exchange price of the coin they mine.

--But how is that possible, Mr. Peabody?  It just don't make sense!
--You see, Sherman, the operators of these "mines" don't make their profits from the coin they mine.  The real money comes from their "investors."  

http://www.ourdailyreflection.com/wp-content/uploads/2012/08/Mr.-Peabody.jpg

It is bad news when BC future depends on investors being robbed of their hard earned money instead of the underlying technology working as intended...

It's working as intended.
--Dad, what's that white stuff in birdshit?
--Son, that's birdshit too.

Where do you live?

So if it is already unprofitable to mine bitcoins, why is the hashrate growing at the constant rate? VCs burning money rushing to increase their network share, but for what reason? Also, if that is the case for some unknown consideration, would not they keep mined bitcoins off the market, does not make any sense to sell at a loss?

Simple overshoot...these devices coming online right now were designed 6 months ago, produced last month....and as soon as they hit the mine they are non-profit. The operators have no choice but to keep going forward...they sunk their costs 6 months ago in the past when the current hashrate was unthinkable.

The reason retail devices are 50% cheaper 2-3 months later is that the mining manufacturers are targeting approximately a 90 day breakeven for their customers. They back calculate that product delivered on such and such a date has a potential in 90 days to produce X bitcoins. So they set the price based on a 90 day break even. If the difficulty is too great, no one makes money (except the manufacturers who sold them 90 days ago).

@Oda, I agree..those two guys are inconsequential in terms of costs, but a wage of some kind has to be on the list of possible expenses. 
The other costs still apply...electricity alone is hundreds of thousands of dollars per month.
Electricity in China is not much cheaper than here in Canada (0.05-0.10 per KW) and energy supply and reliability is a problem in many areas.
Rent may be cheaper....air conditioning may not be needed so badly if the place doesn't have walls. (I've seen that in pics of mines in China.)

Pakistan.

The hashrate is growing because ASIC manufacturers have to unload their miners, which have been in the pipeline for quite a while--since they began designing the chips, which translates to ~8 months on the average.
If you are an ASIC manufacturer, you have two options:

  1.  Eat NRE, pay someone to grind up your stock of chips and haul it away. <==Not smart.
  2.  Sell chips at a loss, partially recoup your investment. <==Pick this if you're smart.
  
If you are a miner, and ordered shit from an ASIC manufacturer months ago, you have two options:

  1.  Sign for the package, walk to the nearest trash can, throw it away. <==Not smart.
  2.  Fire up your miner, and keep mining until it costs more in electricity to mine one coin than the coin's worth.  You won't "ROI," but you'll lose less than with (1). <==Pick this if you're smart.

Re. "keeping mined bitcoin off the market":  Mined bitcoin is no different from any other bitcoin because fungibility.  It makes no more sense to keep mined bitcoin off the market than keeping hoarded bitcoin off the market.

Any of this help?

Edit:
This.

What if all that capital is being put to the wrong use and we have a typical example of tragic malinvestment to be followed by a crash when reality fails to satisfy expectations? This would be the one Bitcoin Bubble deserving of that name.

Nobody said keeping 4000BTC/day off the market would stop the downtrend, the same way new houses kept by the builder can't prevent a housing bubble from busting...

There is always a way to make money through theft.  This is not a Bitcoin phenomenon.

Yes but this can only last for so long. Basic economic principles must apply for the technology to prosper otherwise it crashes towards the real underlying value; in that case minus zero.

Some of you  need to understand, to get lowest operating cost (electrical rate) you need large capital. Add that with getting miner at cost, manufactures are only winners in this game.

Ofcourse they're happy to sell their miners to the public..... WHILE using capital to secure a large farm. They know damn well they will push their customers into red and thus those SOLD MINERS will be off the mining network in short term. Hence, whos winning?

The manufactures only compete with themselves.

I made a chart showing the inflation rate between real BTC inflation and standard BTC inflation (25 BTC/10 minutes).
The ratio is an indicator of the hash rate increase. R=1 no increase, R<1 decrease and R>1 increase.

https://docs.google.com/spreadsheets/d/17Jjhd_nnfRJ9EyYuOM1ACPMwRod3hhPqRk4YQp7FIZE/pubhtml# (last tab)

From November to July, the hash rate decelerate and after start showing signs of acceleration.
The same happened during 2012, after a deceleration until July it started accelerating until november 2013.
At the time the price trend was not so happy. But someone invested in new equipment months earlier to have it online in time to profit from the second big wave.
What they know or think to know we don't know?

And PayPal is teasing about integrating Bitcoin (latest news).

Yeah makes sense pretty much, thanks. So it's all moving basically by inertia, and new production is not being scheduled, and the curve should straighten out somewhere by the end of this year? :) Good, good.

^Lolno.  Of course ASIC manufacturers aren't shuttering--not with all that VC cheddar up for grabs!  And they're gamblers, just like you.  They, and the VC money behind them, are betting on BTC price going up.  
...the only difference being they have a bit more control than you do.  Price spikes?  In-stock gear vanishes, pre-order shipments get delayed, factory-farms ramp up, difficulty spikes until your gear is unprofitable again and then ...and then you get your gear.
*Was a bit creepy, but also sorta' edifying to watch small-time miners (like you?) funding aspiring ASIC manufacturers--and making themselves obsolete in the process.
Because tragedy of the commons.
Because greed.

http://s13.postimg.org/xacj0p8rb/Capture.jpg

A fraction of KNC's farms.

Are those inside the new actic-air-cooled farm or are they located in the old ones? It's mezmerizing to see that amount of mining gear sitting on the shelves and hashing their time away. I wonder what happens to the old miners, it's gotta be a lot of waste...

I thankfully avoided mining (if you don't count CPU mining in 2009) alltogether, but I was very close to ordering some of those vaporware butterflies. Thing that I move around the planet alot saved me from disappointment. But I am waiting for the hash-curve to straighten out because I still want some consumer hashing device that would not make much money but will support the network and make me control a tiny bit of total hashing power. But for such machines to appear and be of apple-quality, meaning polished shelf product, the race has to stop first.

And wow, that picture is impressive.

^Everything keeps going over my head lately--I thought you were being sarcastic :(  Sorry about the digs.
*If the corporate press releases are to be trusted, there's plenty of stuff in the pipeline, and plenty of VC money...