[Proposal] Mining Co-Op

[allinvain]

Sorry for the late reply, I just got back from a brief (well relatively brief for some) vacation. We can easily host 100kw worth of gear. Armedmilitia I replied to your PM. I'm going to discuss your offer with my partner and PM you a rate for 100 kw worth of consumption.

Did you guys decide yet on what miners you will purchase? (SP50 or S7s?)

Right now it seems like we are leaning towards the SP50. That being said it really boils down to price and availability--SP50s are not available at the moment, and we don't know what Spondoolies is charging for them, so I think it is wise for us to wait and see.

It would be wise, or you could just buy S7s now and later on sell them and get SP50s. There would be a slight loss of hashing power and additional funds may be required to get the same equivalent hashing power but you are at least hashing sooner rather than later and thus could also benefit from a potential rise in the btc exchange rate (which likely is an important factor for some of the GB members).

[1714672698]

1714672698

[notlist3d]

Sorry for the late reply, I just got back from a brief (well relatively brief for some) vacation. We can easily host 100kw worth of gear. Armedmilitia I replied to your PM. I'm going to discuss your offer with my partner and PM you a rate for 100 kw worth of consumption.

Did you guys decide yet on what miners you will purchase? (SP50 or S7s?)

Right now it seems like we are leaning towards the SP50. That being said it really boils down to price and availability--SP50s are not available at the moment, and we don't know what Spondoolies is charging for them, so I think it is wise for us to wait and see.

It would be wise, or you could just buy S7s now and later on sell them and get SP50s. There would be a slight loss of hashing power and additional funds may be required to get the same equivalent hashing power but you are at least hashing sooner rather than later and thus could also benefit from a potential rise in the btc exchange rate (which likely is an important factor for some of the GB members).

I think the money is currently not there.  I could be wrong but this seems like a lot of talk of good ideas but no money has been trade hands.  Is there an estimate of how much money will be behind this?

I know some don't think so but I think if find low priced electricity would have a lot more investors.  If you could show .05 or less electricity I think it would not matter S7 or SP50 people would be interested.

[armedmilitia]

The S7 is FAR less power efficient than the SP50. There is an enormous difference between 0.25 and 0.15 w/gh. Unless there is competition in the market (which at the moment, there isn't), it is really not a good idea to buy the S7, as bitmain has no incentive to lower their prices to realistic levels, and their miner has already been eclipsed.

Even at this very optimistic $0.05/kwh, we're looking at 365 days until breaking even on S7s, not factoring in:

• The block reward halving.
• Significant hardware lead time. (best case, they'll be hashing 3-4 weeks from today)
• More efficient hardware being released over the next year which spikes difficulty.
• Cost of PSUs.
• Downtime from damaged miners (as they await RMA), power loss, and internet blackouts.

As for difficulty, the simulation assumes that the rate of difficulty increase stays the same (8.72% per month). That's a very optimistic difficulty estimate, since SP50s haven't even been deployed in large qtys yet.
Sim link: https://coinplorer.com/Hardware/Simulate/56108b61226bf41570cbce49?key=079d11b5-c797-4d29-8f91-61bd6fce8add

Even in the best case, almost fairy-tale like scenario outlined above, our numbers are shit. I am not comfortable in organizing a hardware buy that is doomed by math.

Now, as for the strategy of selling the S7s early and buying SP50's: It's fairly complicated to do that. We'd need to find a buyer/multiple buyers for the s7s, pay escrow fees, hold the funds, negotiate with Spondoolies, and then purchase the SP50s. It exposes us to more risks: What if we can't find a buyer, or we can't negotiate a favourable deal with Spondoolies after selling the miners?

Plus, we still have the problem that we greatly overpaid for our miners, and we are counting on suckers to overpay for ours.

We are in no rush, absolutely no rush at all, to get this done. What we need is competition, price quotes from the competition, and the math to add up. Right now, we have none of those.

[notlist3d]

The S7 is FAR less power efficient than the SP50. There is an enormous difference between 0.25 and 0.15 w/gh. Unless there is competition in the market (which at the moment, there isn't), it is really not a good idea to buy the S7, as bitmain has no incentive to lower their prices to realistic levels, and their miner has already been eclipsed.

Even at this very optimistic $0.05/kwh, we're looking at 365 days until breaking even on S7s, not factoring in:

• The block reward halving.
• Significant hardware lead time. (best case, they'll be hashing 3-4 weeks from today)
• More efficient hardware being released over the next year which spikes difficulty.
• Cost of PSUs.
• Downtime from damaged miners (as they await RMA), power loss, and internet blackouts.

As for difficulty, the simulation assumes that the rate of difficulty increase stays the same (8.72% per month). That's a very optimistic difficulty estimate, since SP50s haven't even been deployed in large qtys yet.
Sim link: https://coinplorer.com/Hardware/Simulate/56108b61226bf41570cbce49?key=079d11b5-c797-4d29-8f91-61bd6fce8add

Even in the best case, almost fairy-tale like scenario outlined above, our numbers are shit. I am not comfortable in organizing a hardware buy that is doomed by math.

Now, as for the strategy of selling the S7s early and buying SP50's: It's fairly complicated to do that. We'd need to find a buyer/multiple buyers for the s7s, pay escrow fees, hold the funds, negotiate with Spondoolies, and then purchase the SP50s. It exposes us to more risks: What if we can't find a buyer, or we can't negotiate a favourable deal with Spondoolies after selling the miners?

Plus, we still have the problem that we greatly overpaid for our miners, and we are counting on suckers to overpay for ours.

We are in no rush, absolutely no rush at all, to get this done. What we need is competition, price quotes from the competition, and the math to add up. Right now, we have none of those.

I am for finding that electricity price of .05 or less the very optimistic price, and not easy to get.   A S7 on .05 electricity costs less then a SP50 on .10 electricity.  This is why I think finding cheap electricity is so important.

On SP50 we really don't know much yet.  We have a picture and stats.  We really have no idea on the release date.  And even once it launches I think it will take multiple to get it at first at least.   So were talking a lot of money.  I could be wrong I just think S7 and SP50 are geared for different customers.   

And I could be wrong I'm just speculating.  I like this idea a lot just think it needs a little tweaking and some things decided eventually.

[Biodom]

1. I am not aware of any datacenter that will charge you only 5c/kwh.
2. We don't know the price of SP50. Without this piece of info, nothing can be calculated.
3. Someone has to show that an actual machine routinely reaches 110 Th and stays there.

[philipma1957]

1. I am not aware of any datacenter that will charge you only 5c/kwh.
2. We don't know the price of SP50. Without this piece of info, nothing can be calculated.
3. Someone has to show that an actual machine routinely reaches 110 Th and stays there.

Not to say I dislike sp-tech  ,    but If I recall  their claim on the sp20 doing 1700gh  would be like me saying my Kia Forte does 120mph.

Yes  my Kia  does  120 mph downhill on a straight dry road in the daylight.


So if you all recall the sp20 really ran well at 1300gh.  about .55watts a gh

And at 1700 more like 1650 gh it did about .75 watts a gh

So it is possible the the sp50 can do 110th  and it is possible it can do .15watts  but that it can not do both at the same time.

My guess is you will want to run the sp50 at  95th in order to see that .15watts

[armedmilitia]

LK Group just posted renders and stats on their upcoming gear. Feast your eyes! We might just be getting the competition I asked for! The more hardware on the market, the better the chances competition will start driving prices down.

Here are details :

Process Node: 14mm
Hashrate Model: 3TH/S , 5TH/S, 10TH/S.. ,
Power consuming : 260W/T
Cooling: heatsink + Dual 12cm Fan
Dimension: 349mm（L）*176mm（W）*148mm（H）
connection port : Ethernet RJ45
Working temperature : 0℃--40℃

[notlist3d]

LK Group just posted renders and stats on their upcoming gear. Feast your eyes! We might just be getting the competition I asked for! The more hardware on the market, the better the chances competition will start driving prices down.

Here are details :

Process Node: 14mm
Hashrate Model: 3TH/S , 5TH/S, 10TH/S.. ,
Power consuming : 260W/T
Cooling: heatsink + Dual 12cm Fan
Dimension: 349mm（L）*176mm（W）*148mm（H）
connection port : Ethernet RJ45
Working temperature : 0℃--40℃

LKETC has as long history with it's dragon's (or A1's by technical name).  They have been a big producer in the past.  They partner with Innosilicon on chips.  Only thing is they announced kinda early and now did a reminder.

As far as timeline it's more to do with Innosilicon.  Once they get chips going no doubt LKETC will be partners again.  But when that will be is the question.

[QuintLeo]

I am for finding that electricity price of .05 or less the very optimistic price, and not easy to get.

 Trivial to get in the US - *IF* you live in one of 3 counties in Washington State. Pretty much impossible about anywhere else in the US AFIAK.

There is an enormous difference between 0.25 and 0.15 w/gh.

 But can the SP50 hit 110TH at 0.15 w/gh? Given Spondoolies record, I'd guess that the 2 figures are mutually exclusive - pick one or the other but not both.
 I'd guess closer to 80TH to hit that 0.15 figure, but 'till actuall hardware hits the field it's hard to tell.

 I suspect the 8.72%/month of the "simulation" might be a bit pessimistic - but I wouldn't bet on much less than 8% (4% per difficulty adjustment appx) for the rest of this year and well into next year. Too many announcements of "new hardware" yet current hardware is still profitable (SP20/30/possibly other Spondoolies and S5s should stay profitable 'till the halfing, though they might just be BARELY so by June 2016).

By simply using a meer 700kwh or more per month the rate can be around $.08-.09/kwh depending on your area.

 Standard residential rate around here *EXCEPT for 3 months in the summer*, I get appx. 0.67 (all charges and tax included) for every KWH past about 1200 per month. There is also a "Time of Day" rate that doesn't save much on that last "step" in the rate, but does cut a noticeable amount from the first 2 steps of the rate.

 A single S5 will go through about 430 KWH a month, so it's NOT hard to get to the top tier of the residential rates here.

 On the other hand, June 16 through September 15 at around 14.2c/KWH (with ALL adjustments AND ToD rate included and NO lower tiers with higher usage) is pretty bad.


 I've not looked at commercial rates around here at all, since I have more power than I can use at home.

[notlist3d]

Lower rates are possible look over at Mr. Lees thread like one away from this:

we negotiate with a new electricity supplier those day, if we achieve agreement, the host fee will deduce to 0.07$ per kw.h on this new farm.
however, we want to do another host model in this farm, eg cloud mining mode.

we want to sell hash rate directly, the price should be 369$ each Th/s @ 0.3watt/gh/s , so the daily host fee will be 0.3*24*0.07=0.504$ for each Th/s,  we will calculate the mining earning according to 4%pps and pay the earning to customer each  day after deduce daily host fee(we will transfer daily host fee to bitcoin according to exchange ratio from bitfinex at 9:00AM(UMT+8)each day).

that is a similar model like hashnest, the only difference is that our host fee is better, and i also will not charge for expensive off-shelf fee when the project is end. i will try to sell the miner on used miner market and pay the money directly to customer.

He is most likely going to manage .07 well before hashing.  This is greater then any offer I think we have been given.    If we get our ducks in a row I still say electricity price is a huge factor we should focus on.

The big winners will have cheap electricity.   

[GreatNorthData]

Lower rates are possible look over at Mr. Lees thread like one away from this:

we negotiate with a new electricity supplier those day, if we achieve agreement, the host fee will deduce to 0.07$ per kw.h on this new farm.
however, we want to do another host model in this farm, eg cloud mining mode.

we want to sell hash rate directly, the price should be 369$ each Th/s @ 0.3watt/gh/s , so the daily host fee will be 0.3*24*0.07=0.504$ for each Th/s,  we will calculate the mining earning according to 4%pps and pay the earning to customer each  day after deduce daily host fee(we will transfer daily host fee to bitcoin according to exchange ratio from bitfinex at 9:00AM(UMT+8)each day).

that is a similar model like hashnest, the only difference is that our host fee is better, and i also will not charge for expensive off-shelf fee when the project is end. i will try to sell the miner on used miner market and pay the money directly to customer.

He is most likely going to manage .07 well before hashing.  This is greater then any offer I think we have been given.    If we get our ducks in a row I still say electricity price is a huge factor we should focus on.

The big winners will have cheap electricity.   

I think I offered 0.07 if you have at least 100kW

[notlist3d]

Lower rates are possible look over at Mr. Lees thread like one away from this:

we negotiate with a new electricity supplier those day, if we achieve agreement, the host fee will deduce to 0.07$ per kw.h on this new farm.
however, we want to do another host model in this farm, eg cloud mining mode.

we want to sell hash rate directly, the price should be 369$ each Th/s @ 0.3watt/gh/s , so the daily host fee will be 0.3*24*0.07=0.504$ for each Th/s,  we will calculate the mining earning according to 4%pps and pay the earning to customer each  day after deduce daily host fee(we will transfer daily host fee to bitcoin according to exchange ratio from bitfinex at 9:00AM(UMT+8)each day).

that is a similar model like hashnest, the only difference is that our host fee is better, and i also will not charge for expensive off-shelf fee when the project is end. i will try to sell the miner on used miner market and pay the money directly to customer.

He is most likely going to manage .07 well before hashing.  This is greater then any offer I think we have been given.    If we get our ducks in a row I still say electricity price is a huge factor we should focus on.

The big winners will have cheap electricity.   

I think I offered 0.07 if you have at least 100kW

I don't know that .07 was offered that is a little more promising then I though.  You were more like this:

If there was a big group buy of hardware sent to you could you do as far as lowest price that is your next drop?  And when would you expect it to happen?

There are so many factors. I don't want to over promise and under deliver. Rates are going to continue to drop for the foreseeable future, but what I've posted in this thread is the best we can do right now. The answer could very well be different in a few months.

I still think magic number to try to get is .05.   I feel data centers are finding better and better places and prices are shrinking.  So I'm hoping around there still.

But I can understand not giving exact numbers as this CO-OP has been pretty slow moving.  I still have hope it will speed up once a new product hits market.

[Searing]

LK Group just posted renders and stats on their upcoming gear. Feast your eyes! We might just be getting the competition I asked for! The more hardware on the market, the better the chances competition will start driving prices down.

Here are details :

Process Node: 14mm
Hashrate Model: 3TH/S , 5TH/S, 10TH/S.. ,
Power consuming : 260W/T
Cooling: heatsink + Dual 12cm Fan
Dimension: 349mm（L）*176mm（W）*148mm（H）
connection port : Ethernet RJ45
Working temperature : 0℃--40℃

Anyone on here have an educated guess on what say the 3TH above would go for? or the 5TH admitedly this would be I suppose by the time we
would see such a unit be Jan or Feb 2016 at my guess

Just curious I'm very bad at the above price guessing game. I'm always low.

[notlist3d]

LK Group just posted renders and stats on their upcoming gear. Feast your eyes! We might just be getting the competition I asked for! The more hardware on the market, the better the chances competition will start driving prices down.

Here are details :

Process Node: 14mm
Hashrate Model: 3TH/S , 5TH/S, 10TH/S.. ,
Power consuming : 260W/T
Cooling: heatsink + Dual 12cm Fan
Dimension: 349mm（L）*176mm（W）*148mm（H）
connection port : Ethernet RJ45
Working temperature : 0℃--40℃

Anyone on here have an educated guess on what say the 3TH above would go for? or the 5TH admitedly this would be I suppose by the time we
would see such a unit be Jan or Feb 2016 at my guess

Just curious I'm very bad at the above price guessing game. I'm always low.

Honestly there is no way to know.  A1 dragons originally went for over 3k.  But dropped and dropped.... etc till they are in several hundreds today in many cases.

I don't think it will be near the first A1 dragons price.   But I don't see them playing a big game with bitmain on trying to be cheaper then them.  I suspect a tad cheaper then what ever bitmain is selling at the time.   But they will want profit so no drastic changes from bitmain prices I dont think.

Only company to really force bitmain to drop price in war almost was SP.  And they claimed on some of last sp20's they lost money not sure if true or not. 

[Searing]

LK Group just posted renders and stats on their upcoming gear. Feast your eyes! We might just be getting the competition I asked for! The more hardware on the market, the better the chances competition will start driving prices down.

Here are details :

Process Node: 14mm
Hashrate Model: 3TH/S , 5TH/S, 10TH/S.. ,
Power consuming : 260W/T
Cooling: heatsink + Dual 12cm Fan
Dimension: 349mm（L）*176mm（W）*148mm（H）
connection port : Ethernet RJ45
Working temperature : 0℃--40℃

Anyone on here have an educated guess on what say the 3TH above would go for? or the 5TH admitedly this would be I suppose by the time we
would see such a unit be Jan or Feb 2016 at my guess

Just curious I'm very bad at the above price guessing game. I'm always low.

Honestly there is no way to know.  A1 dragons originally went for over 3k.  But dropped and dropped.... etc till they are in several hundreds today in many cases.

I don't think it will be near the first A1 dragons price.   But I don't see them playing a big game with bitmain on trying to be cheaper then them.  I suspect a tad cheaper then what ever bitmain is selling at the time.   But they will want profit so no drastic changes from bitmain prices I don't think.

Only company to really force bitmain to drop price in war almost was SP.  And they claimed on some of last sp20's they lost money not sure if true or not. 

Well I'm in the USA with a 13.5c usd kwh rate. So the whole question is likely MOOT. With bitfury announcing the 100 million data hall expansion and the sp50's 110TH
units and additional data hall by spondoolies and of course knc expanding its private chip data hall with IPO $$$

well hell them big boys could find themselves with a major diff war and the machines may only last 4-5 months before they supersize difficulty too the moon (if not price..it is btc by thw way it only loves 'sideways')

So best I'm out....too many toys ..too cheap elec around the world..used/new/data halls it is like "Go east young man"

Bitcoin always a drama queen!

[QuintLeo]

They might have a market for the 10TH unit at $3000 - no way they'll make any sales if the 5TH or 3Th units are even close to that.

Given the watts/TH are almost the same as the S7, they really need to at LEAST match S7 pricing pretty closely, and given they're comming to market later they really need to beat S7 pricing per TH by a bit.

[Searing]

They might have a market for the 10TH unit at $3000 - no way they'll make any sales if the 5TH or 3Th units are even close to that.

Given the watts/TH are almost the same as the S7, they really need to at LEAST match S7 pricing pretty closely, and given they're comming to market later they really need to beat S7 pricing per TH by a bit.

Yeah but after watching how this stuff works (equip/vs difficulty) by the time such units MAY come out it will be Jan/Feb imho and by then bitfury's 100 million buck data hall is up and so is the large sp50 farm with whatever they allow to go outside as bulk spondoolies that is

and of course it will prob be 6k for such rather then 3k for such and you'd be no better off then an S7 now

seems how that always shakes out (and don't forget most times equip is late by at least a month (spondoolies) or 4-5 months (knc) even their private 16nm chip was late 5 months from what i can tell (knc) for their ipo farm

yeah diff is gonna go way way up with all them toys sitting in data halls hitting the world imho

but yeah 3k I'd at least be tempted like the S7 but again in my situation at 13.5c kwh the S7 is pretty to look at but really ..and only a 90 day warranty? If a controller
board breaks you have to RMA all the way back to china? no pay up front and ship and swap the broken component back with rma for credit when it arrives (they'd kill if they did that)

naw too many risks for me even w/o the elec costs....ie again the stuff always seems to arrive months late or on time with issues and my friend this coming year difficulty will wait for NO man ..to pun a famous phrase:

(sits in bleachers eating peanuts watching this asic soap opera)

in that this is the mining co-op thread their MAY be some hope for group buy or such a setup working (to be fair) individually not so much imho.

[carlosmnk]

I am surprised a 14nm have the same eficency than a 28nm... 0.26w/ghs? really? i ever thought a 14/16nm will be on 0.06w/ghs o 0.08w/ghs and above... up to 0.01w/ghs 
But it will help us to buy cheaper... 

[notlist3d]

I am surprised a 14nm have the same eficency than a 28nm... 0.26w/ghs? really? i ever thought a 14/16nm will be on 0.06w/ghs o 0.08w/ghs and above... up to 0.01w/ghs 
But it will help us to buy cheaper... 

The company's that kept 28 NM really did a great job getting more out of it.  Right now BM's S7 they pushed it pretty well.  And if SP50 comes like spec they got even better out of their chip.

I think the ones that kept current NM and perfected it got a head start and easier R/D.   Them going to 14  NM might be good for long term.  But the short term it left them with no miners to sell and a lot longer R/D

[ATCkit]

Count me in. I've done a ton of research on hosting in the past. GND is the best option I had seen the last time I checked.

As far as adding more "hashing power" to this group buy, i think you guys could team up with the Get Hashing group and be able to get larger numbers of people involved. Get Hashing has been doing what you guys are wanting to do for some time. They're on their 3rd batch now. Only difference is they have their data center in the US. Canada, GND or another CDN doing the same thing, will help to get the cost down further.

In a nutshell, it's all about rig efficiency and lowest cost for hosting. The clear choice at this point appears to be:

- the SP50  (Maybe an even more efficient rig will be available by Q1 2016.)

- hosted at GND.

- Get Hashing groups Proof of Everything and their trusted/known pay out model. (perhaps Armed or someone could talk to cyberdexter, keymaster, daffy or nemesio about this?)

Having said all that, none of us can tumble the numbers until we get a price for the SP50 or whatever may be competitive with it in Q1 2016. However, I like the way you guys are thinking.


EDIT: Here's the Get Hashing forum link: https://forum.gethashing.com/  and
how they're already handling dispersal of batch 1 units as they move into batch 3 can be found here: https://forum.gethashing.com/t/ann-ghx-batch-3-hardware-redemption-round-1/5485

Note: I'm not trying to hi jack this group buy idea. I think that like minded people can work together and make it a win win. Get Hashing already has a lot of trust built up with their Proof of Everything (POE) concept.

EDIT 2: Sorry if i missed this. Looks like allinvain has a new site -Cryptoboreas in Montreal. Great prices too. It simply makes sense that Quebec and NFLD/Labrabor have the lowest costs globally.