CoinTerra Update: Engineering and Production status

[testerx]

The production units that we have assembled and tested so far are running at between 1.63TH/s and 1.72TH/s with power draw at the wall between 1900W and 2100W – typically 20% higher than our anticipated target of 1650W

20%??? What happened to:

The chip will consume less than 300w, for a performance of 504 Gh/s.

Sounds like you missed the power target by quite a bit more than 20%.

they didnt say what their asics draw.  they said what the system draws at the wall.. which includes a built in controller, plenty of fans, a liquid cooling system (pumps).  and they confirmed that their power conversion circuitry is what theyre redesigning for the new version...

anyway... what any asic company says (based on simulation) before they get their chips back, and what they say (when its fully tested and includes all the real-world losses) after their chips are in a system always seem to be different.. and i cant think of anyone who delivered a system with the same power draw as their specs.   can you?

WOW, my legitimate concerns have been confirmed now, and you have been proved to be on the wrong side of this debate...yet you continue to defend CT despite their clear failures.  Are you a shill?  Or just defending a position you've over-invested yourself in emotionally?

He's one of their angel investors, I don't know why you would attack him for feeling invested, he's literally invested.

[Cheshyr]

You really think difficulty would be over 20 billion in May?  Even with hashfast and cointerra shipping that is somewhat unlikely, 10 billion is more realistic in the spring.  If you really think difficulty is 22 billion in May then nobody would be buying KnC units at all.

The problem isn't that these won't ROI, it's that these now require huge power infrastructure, it's not cheap or easy to come up with the kind of power solutions needed to run these and for people who've basically sent all their spare cash to cointerra it's not exactly easy to come up with tons more money to build a setup capable of running crazy 2000 watt sucking rigs.  And if you've already invested in new power lines/PDUs/cabinets/etc it might be a loss.  

I have no idea what the difficulty will be... I was using the default numbers from that site to make a point that shipping now was in the customers best interest, even with the excessive power usage and underperformance.  It's definitely non-ideal...  but losing almost $8k in mining revenue while they attempted to fix it would also be non-ideal.  This is still the lesser of two evils, and CoinTerra acted appropriately in the light of that reality.  That was all.

[aerobatic]

WOW, my legitimate concerns have been confirmed now, and you have been proved to be on the wrong side of this debate...yet you continue to defend CT despite their clear failures.  Are you a shill?  Or just defending a position you've over-invested yourself in emotionally?

really?  someone doesn't agree with you and they're a shill?

you can't possibly be pragmatic and think of the best longterm value proposition, or is it better to be right, regardless of how much value you extract?

im not on the wrong side on 'the debate'  there's a debate - there is no 'wrong side' - and i try to be balanced and consider both sides of any argument.  

In this case I'm trying to focus on the big picture.  How to maximise value from our mining gear.

im not defending CT.  I'm waiting for my boxes just like you.  AND i want to extract maximum value out of my boxes - which means getting them sooner, and is much more important to me than whether they run at full speed or with a 15-20% under clock

Do you agree that the difficulty rises by about 20-30% each level change?  Thus you know that every week of delay in shipping, has a lost value to us equal to the diff change?

So.. what would you prefer that cointerra do - when faced with this situation?   ship boxes as fast as they can, whatever speed they run?... or take their time tweaking it, and deliver it when it can hit 2 TH - even if that could be 8 weeks away?

actually. lets put this in terms that are easier to compare...   do we prefer CT try to be more like KNC... and deliver quickly to customers... or more like HF, who are tweaking indefinitely and seem in no hurry to ship to customers ?    since the lateness of the hardware has a bigger impact on its ability to mine bitcoins, than exactly how fast it runs... don't we all agree that shipping the Jan customers sooner at 1.6 TH/s is Wayyy better than shipping later, at 2 TH/s?

Which one do you prefer?   I know which i prefer...  i want my TerraMiner BOX, NOW... even if it only hashes at 1.6 TH/s box - rather than wait 8 weeks for a 2 TH/s box.   Sure, the later batch customers will get a faster box than i will.  but i will have received my boxes sooner, and will have had a bigger benefit from mining earlier.

and yes, i want compensation too... but i care about the delivery of my box quickly much more than i care about whether I'm going to get a bonus or refund or discount or whatever they decide to do.  One week earlier shipping, more than covers any nickel or dime i could squeeze out of them in refunds or other comps.

As for how shilly i am... my record speaks for itself.. I've made a few hundred posts.   you can see how biassed or balanced i am.   and also, i'm not an anonymous bitointalk handle...  you know who i am.   perhaps i should also remind that i've got just as big - actually, bigger order with hashfast as i do with cointerra.  and i am also a knc customer... and bitfury... and bfl.. and previously also asicminer investor.  in short, I've nothing to gain from shilling... all i want is to mine productively and earn as many coins as possible.  if cointerra has good hardware for that so be it.   currently, I've been mining with knc jupiters and they've done well for me and I'm extremely happy with them as a company, and the boxes themselves.

i try and spread my eggs, and operate a portfolio approach to my mining gear.  i try not to pick sides.. and try to de-risk my purchases by ordering a bunch from several of the asic companies at the same time.  but i don't order from all of them, because i don't always agree with some of their technical or business decisions.  but that doesn't mean i go around dissing them either.

[bronxnua]

20% missed power target. 2 circuits instead of 1. You can still probably get away with using 20A circuit @ average US home.

I remember reading the post here when you guys announced the chip for a first time. Glad I didn't fall for that. Same goes to hashfast.

Nope. trying to pull 2000w from a single US 120v is a terrible idea.

there are two independent power supplies.   

so if you have 2 circuits that would be okay.  I did that in anticipation. I have 2 20 amp circuits waiting.

[hardpick]

If they shipped the expected 2000GHs now, started mining in Feb '14, and took the default assumptions on the website, you'd be looking at a lifetime cumulative return of $9190.  I know, this is a completely fictional number, but we have to start somewhere.  http://thegenesisblock.com/mining/a/ac1fbf2676

Sadly, reality happened.  They are shipping 1800GHs now, so you can start mining in Feb '14.  Lifetime cumulative return of $7510.  http://thegenesisblock.com/mining/a/9090810af2

You're down about $1700, but you're still very profitable.

Batch 1 might start mining Feb '14. Those cost $15,750 and are thus going to be a huge loss due to the delay. Who knows when Batch 2 will ship.

Ouch.  My bad; those are batch 1 units?

$14k, 2TH, 1.6kW... $1190 lifetime cumulative return if started in Feb '14.  http://thegenesisblock.com/mining/a/cafe5bf642
$14k, 1.8TH, 1.6kW... -$460 lifetime cumulative return if started in Feb '14.  http://thegenesisblock.com/mining/a/30119efaf4
$14k, 2TH, 1.6kW... -$6000 lifetime cumulative return if started in Mar '14.  http://thegenesisblock.com/mining/a/e5232335d0

So, point still stands; CoinTerra is minimizing customer losses by shipping early at customer request.  That said, $14k for batch 1...  wow.

You really think difficulty would be over 20 billion in May?  Even with hashfast and cointerra shipping that is somewhat unlikely, 10 billion is more realistic in the spring.  If you really think difficulty is 22 billion in May then nobody would be buying KnC units at all.

The problem isn't that these won't ROI, it's that these now require huge power infrastructure, it's not cheap or easy to come up with the kind of power solutions needed to run these and for people who've basically sent all their spare cash to cointerra it's not exactly easy to come up with tons more money to build a setup capable of running crazy 2000 watt sucking rigs.  And if you've already invested in new power lines/PDUs/cabinets/etc it might be a loss.  

"22 billion in May"


is very possible  at present 2% per day compound growth = 22 billion at end of may / early june ( this has being the trend for about 8 months)
it possible Neptune will ship by mid May
shipping soon  = bfl/hashfast/cointerra and others

[Syke]

actually. lets put this in terms that are easier to compare...   do we prefer CT try to be more like KNC... and deliver quickly to customers...

Did you really just compare CT with KNC? KNC delivered 2 weeks late with 50% greater hashrate. Customers were fairly compensated.

CT is delivering 2 months late, at -20% hashrate. Ouch!!!

[miahallen]

im not defending CT.

Really 

really?  at 1.6 TH in a 4U single box, you're a bit disappointed ?   they just showed proof... working... the fastest, lowest cost, and smallest working bitcoin mining box that has ever existed... running with only four chips and by the looks of it, at 70 degrees according to their stats, they're not even breaking a sweat....  and you're disappointed?     And they said each day their engineers make it go faster...

they've literally gone from yesterday not being in the mining business... to today, having an extremely impressive box...  hashing faster than any box I've ever seen... that i would happily take delivery of tomorrow...!

Now all they're doing is tweaking it to go 20% faster before they ship it?  heck, thats the smallest leap of faith I've ever been asked to make.

more here:
https://bitcointalk.org/index.php?topic=416576.20

I agree with you that at this point, the best thing to do is to deliver what they have.  My complaint is the same one we discussed a few weeks ago, that they have over-promised and under-delivered.  This doesn't seem to bother you, and I'm trying to figure out why.

[aerobatic]

I agree with you that at this point, the best thing to do is to deliver what they have.  My complaint is the same one we discussed a few weeks ago, that they have over-promised and under-delivered.  This doesn't seem to bother you, and I'm trying to figure out why.

course it 'bothers me' that i will receive 15% less hash power than i was expecting.  But in the time since i ordered, bitcoins are worth 800% more than they were when i made the decision to buy...   so I'm significantly better off than i forecasted i would be...!  (and the hardware is consequently worth a lot more... if it were sold today!)   bitcoin mining hardware value correlates well with price of bitcoin.

but 'bother' is the wrong word.  this is business.  and my portfolio is diversified.  mining is one of my areas of interest and i invest in other assets (properties, companies etc).  

I've said this before.. but i'm less worried about taking delivery of my CT gear within a few weeks of schedule (at 15% less than expected hash rate) than i am at not receiving my HF gear, 12+ weeks later than scheduled... (regardless of whatever hash rate hf will over-delver).  and i paid more for my hf gear than my ct gear, $/gh.   bottom line, the delivery date is more important (read, more valuable!) than the hash rate.

[IamCANADIAN013]

So I can pick one up (TerraMiner™ IV – 2TH/s, Networked ASIC Miner) for $6,391.97 U.S. to be delivered in May.

For the price, this seems to be the best bang for your buck that I have come across so far.

The only big variable i see is mining difficulty. Am I missing something?

I notice they only take Bitcoin and bank transfer, that's a big red flag to me, or is this normal with companies like this?

Do any take credit cards?

[Syke]

So I can pick one up (TerraMiner™ IV – 2TH/s, Networked ASIC Miner) for $6,391.97 U.S. to be delivered in May.

For the price, this seems to be the best bang for your buck that I have come across so far.

The only big variable i see is mining difficulty. Am I missing something?

I notice they only take Bitcoin and bank transfer, that's a big red flag to me, or is this normal with companies like this?

Do any take credit cards?

That's at least 3 months away, difficulty will go up 10x in that time. So 200 GH/s now is equal to 2 TH/s in May. You can get 200 GH/s now for just over $1000. Why would you throw away $5000???

[IamCANADIAN013]

So I can pick one up (TerraMiner™ IV – 2TH/s, Networked ASIC Miner) for $6,391.97 U.S. to be delivered in May.

For the price, this seems to be the best bang for your buck that I have come across so far.

The only big variable i see is mining difficulty. Am I missing something?

I notice they only take Bitcoin and bank transfer, that's a big red flag to me, or is this normal with companies like this?

Do any take credit cards?

That's at least 3 months away, difficulty will go up 10x in that time. So 200 GH/s now is equal to 2 TH/s in May. You can get 200 GH/s now for just over $1000. Why would you throw away $5000???

200 gh/s for $1000?  I haven't been able to find anything close to that.  I'll have to keep looking.

[aerobatic]

So I can pick one up (TerraMiner™ IV – 2TH/s, Networked ASIC Miner) for $6,391.97 U.S. to be delivered in May.

For the price, this seems to be the best bang for your buck that I have come across so far.

The only big variable i see is mining difficulty. Am I missing something?

I notice they only take Bitcoin and bank transfer, that's a big red flag to me, or is this normal with companies like this?

Do any take credit cards?

That's at least 3 months away, difficulty will go up 10x in that time. So 200 GH/s now is equal to 2 TH/s in May. You can get 200 GH/s now for just over $1000. Why would you throw away $5000???

difficulty probably won't go up 10x in 3 months.  thats quite a pessimistic view.

we're at end of jan and the network is 17.5 PH today.  So you think think it could be seriously be 175 PH by end of April ?

And, as we know.. the lowest cost bitcoin mining equipment available today costs $3/GH and some of it is trading at $20-40 per GH, so that $3/gh price is at the extreme low end of the market ...  so for the sake of argument, lets assume that it stays at that price for the next 3 months...

So you're saying that with 175 - 17.5 PH.. thus 157 PH will be sold (or has been sold) to be deployed on the network in the next 3 months.

157 PH, at $3/GH (or, $3M per PH), is $472.5m's worth of bitcoin mining equipment.

you really think bitcoin miners will spend and deploy nearly $500m on bitcoin mining hardware in the next 3 months?

then lets compare electricity usage...

The best bitcoin miners today, and also anything available in the next 3 months, is approximately 1 watt per GH.. which equates to 1 Megawatt per Petahash.

Thus you are saying that 157 Megawatts of additional continuous power will be required for bitcoin mining in the next 3 months?

i really think thats unlikely.  i don't think you can simply take what bitcoin mining has done over the last few months and extrapolate even the next few months worth of use as we've undergone a transition from slow asics to fast asics, and that just can't continue at the same rate.  sure, it will keep growing, but at a slower growth rate.

-- Jez

[gmaxwell]

And, as we know.. the lowest cost bitcoin mining equipment available today costs $3/GH and some of it is trading at $20-40 per GH, so that $3/gh price is at the extreme low end of the market ...  so for the sake of argument, lets assume that it stays at that price for the next 3 months...

Don't confuse retail with "cost", do you think KNC is paying $3/GH for the hashrate they are putting online themselves?

I agree that 10x in three months is pessimistic, but 6x is what 2%/day suggests— and thats what we've been averaging for some time now (http://bitcoin.sipa.be/growth.png).  It can't continue forever, of course, but long enough to prevent profit for everyone buying at $3/GH delivered months from now? I absolutely think it can.

[aerobatic]

And, as we know.. the lowest cost bitcoin mining equipment available today costs $3/GH and some of it is trading at $20-40 per GH, so that $3/gh price is at the extreme low end of the market ...  so for the sake of argument, lets assume that it stays at that price for the next 3 months...

Don't confuse retail with "cost", do you think KNC is paying $3/GH for the hashrate they are putting online themselves?

I agree that 10x in three months is pessimistic, but 6x is what 2%/day suggests— and thats what we've been averaging for some time now (http://bitcoin.sipa.be/growth.png).  It can't continue forever, of course, but long enough to prevent profit for everyone buying at $3/GH delivered months from now? I absolutely think it can.

ok.. reworking some of the assumptions...   if asic companies made their own asics and system boards and deployed their own asics in their own private mines...  they could buy the systems cheaper than retail, of that there is no doubt.   but the costs aren't zero.  its at best (in this generation) half the retail cost.

i suspect the cost price of a knc jupiter s probably still approx half the retail cost, and don't forget that the $3/gh that I've quoted is a cointerra price (i.e. cheapest that i currently know of), not a knc price (which was $17/gh for the 1st jupiters in oct).. later jupiters were cheaper and faster is still probably in the $8-$10/gh range.

if you're thinking that knc can make a jupiter (for their own in-house use) for $1/gh, i think thats much too low and not achievable in the current generation.  of course, with 20nm maybe its closer to reality.   And by comparison, look at the new asicminer chips announced yesterday, in the $1/gh range.. just for the asics.  when you add boards, cooling, power, controller, etc... the price is going to be north of $1.5-$2/gh.

and think about how many millions you need to devote to build out your private mine.   21e6 raised $5m at a 6m pre-money valuation...   what does $5m buy you?   maybe a full asic design and a first batch of production.  so maybe they could afford the first PH at most.  they can't afford any more without raising more cash, or mining to raise the cash, to build up the mine further.

and why would they do it when there are plenty of people willing to give them cash, now, to sell their miners.. cash in advance is much lower risk and a better return for them than mining, where they take on network diff risk and deferred cash-flow.

[joeventura]

Congratulations to the CoinTerra team! I know how exciting it is to send your new baby out the door for the first time! Good work guys (and gals if there are any!).

JINX!!

[Syke]

That's at least 3 months away, difficulty will go up 10x in that time. So 200 GH/s now is equal to 2 TH/s in May. You can get 200 GH/s now for just over $1000. Why would you throw away $5000???

difficulty probably won't go up 10x in 3 months.  thats quite a pessimistic view.

we're at end of jan and the network is 17.5 PH today.  So you think think it could be seriously be 175 PH by end of April ?

Way to miss the point. CT is highly overpriced. CT = $3/GH, Bitmain = $6/GH. The difficulty is going up way more than 2x in 3 months. I guarantee it. Dare to bet on it?

[Unacceptable]

Congratulations to the CoinTerra team! I know how exciting it is to send your new baby out the door for the first time! Good work guys (and gals if there are any!).

JINX!!

Thanks for the laugh joe !!!!!!!!!!!!!!!!   

[jermwerty]

200 gh/s for $1000?  I haven't been able to find anything close to that.  I'll have to keep looking.

https://bitmaintech.com/productDetail.htm?pid=00020140107162747992Ce5uBuxW06D6

[aerobatic]

200 gh/s for $1000?  I haven't been able to find anything close to that.  I'll have to keep looking.

https://bitmaintech.com/productDetail.htm?pid=00020140107162747992Ce5uBuxW06D6

its a different product category.  the antminer looks like its aimed at the hobbyist end of the market (same as bitfury).. it doesn't seem quite as professional as some other products...  it doesn't have a case, nor a power supply, and seems less 'plug and play' (and costs more, and uses more power per gh, etc).  but yes, its available from stock, and that counts for a lot.

if i didn't already have an early order with knc, hashfast and cointerra i would've been more interested.  But I'm eagerly awaiting the wasp to see how they get on.

-- Jez

[IamCANADIAN013]

200 gh/s for $1000?  I haven't been able to find anything close to that.  I'll have to keep looking.

https://bitmaintech.com/productDetail.htm?pid=00020140107162747992Ce5uBuxW06D6

Thanks for the link.

I'd still be looking at roughly $1340 Canadian, plus shipping, can't see how much they charge, I'm assuming it isn't free.

Then I'm gonna get nailed by customs for tax and duty I assume. Tax alone would be about $160. There are people listing these on Craigslist for $3500 here though.

Maybe some money in just reselling them, buy bitcoin with that.