ASICMINER: Entering the Future of ASIC Mining by Inventing It

[Vycid]

Care to say which of those 3 is derived from the number being calculated? It is most certainly not clear. Are you confusing sale price and manufacturing cost?

Sale profits.

https://bitcointalk.org/index.php?topic=99497.msg2788821#msg2788821

Using the sale price of the hardware (which was known), it is possible to determine how many blades were manufactured for sale based on the sale proceeds. Hardware deployed is known: 50 TH (FC did disclose this number, look through his post history). Summing these two and dividing by manufacturing cost gets you a cost per hash figure.

I have no interest in finding the scratch paper where I did this nor in typing it up. Do it yourself.

As mentioned, USBs are a confounding variable and require some guesswork. If their manufacturing cost was significant, then $1.50 is too low an estimate, which I am obviously comfortable with.

[101111]

Right. Intelligent conversation like the pages and pages of "TO THE MOON!" and "when are we going to hit 10 BTC??", followed by "WHY DID THE SHARE PRICE DROP???"

The appearance of a contrarian opinion was a warning sign to those who weren't keeping their heads in the sand.

More lies easily disproved with a simple search.

[Vexual]

https://bitcointalk.org/index.php?topic=99497.msg2788821#msg2788821

Using the sale price of the hardware (which was known), it is possible to determine how many blades were manufactured for sale based on the sale proceeds. Hardware deployed is known: 50 TH (FC did disclose this number, look through his post history). Summing these two and dividing by manufacturing cost gets you a cost per hash figure.

I have no interest in finding the scratch paper where I did this nor in typing it up. Do it yourself.

As mentioned, USBs are a confounding variable and require some guesswork. If their manufacturing cost was significant, then $1.50 is too low an estimate, which I am obviously comfortable with.

So do we take mining income as displayed on Friedcat's July 23 Update as being profit after expenses too?
You have no way of relating expenses listed on that update to the expenses directly related to the hardware sold over that period.
You also dont know what price some of the hardware was sold at, for example, cost to resellers and direct bulk buys.
This 1.50/GH number that your calculations rely on could be out by some order of magnitude.

[Vycid]

https://bitcointalk.org/index.php?topic=99497.msg2788821#msg2788821

Using the sale price of the hardware (which was known), it is possible to determine how many blades were manufactured for sale based on the sale proceeds. Hardware deployed is known: 50 TH (FC did disclose this number, look through his post history). Summing these two and dividing by manufacturing cost gets you a cost per hash figure.

I have no interest in finding the scratch paper where I did this nor in typing it up. Do it yourself.

As mentioned, USBs are a confounding variable and require some guesswork. If their manufacturing cost was significant, then $1.50 is too low an estimate, which I am obviously comfortable with.

So do we take mining income as displayed on Friedcat's July 23 Update as being profit after expenses too?
You have no way of relating expenses listed on that update to the expenses directly related to the hardware sold over that period.
You also dont know what price some of the hardware was sold at, for example, cost to resellers and direct bulk buys.
This 1.50/GH number that your calculations rely on could be out by some order of magnitude.

Obviously mining income is ignored, we already know the size of the farm. We are only interested in the total hashrate produced in order to determine a cost per gigahash.

It could be off, but guess what? All this would make per-GH costs higher, not lower.

I've established a best-case scenario for AM and it STILL appears significantly overvalued. Thus my short position.

[Vexual]

It could be, but guess what? All this would make per-GH costs higher, not lower.

I've established a best-case scenario for AM and it STILL appears significantly overvalued. Thus my short position.

Cool for you, but your maths is wrong everywhere, even at $15/GH profit margin is better than 75%

[Vycid]

It could be, but guess what? All this would make per-GH costs higher, not lower.

I've established a best-case scenario for AM and it STILL appears significantly overvalued. Thus my short position.

Cool for you, but your maths is wrong everywhere, even at $15/GH profit margin is better than 75%

You fundamentally misunderstand. That 75% is the average projected over the next year.

If Cointerra is selling at $3/GH, do you see the issue with AM producing at $15?

Even if AM's margins stay near 100 for the near future, if they drop below 50 later in the year the average could easily dip below 75%.

[Vexual]

It could be, but guess what? All this would make per-GH costs higher, not lower.

I've established a best-case scenario for AM and it STILL appears significantly overvalued. Thus my short position.

Cool for you, but your maths is wrong everywhere, even at $15/GH profit margin is better than 75%

You fundamentally misunderstand. That 75% is the average projected over the next year.

If Cointerra is selling at $3/GH, do you see the issue with AM producing at $15?

No, I just wanted to see how you'd react to something completely made up.

[freedomno1]

Nope real number is $52/GH based on old gen I assume Vycid just put in a random number for the next gens cointerra is 3/GH though if we assume delivery in January 2013 in my opinion cointerra being labeled better than a BFL monarach on price per GH is kind of suspicious.

If that is true, ASICMiner is probably worth less than 0.5 BTC/share. They'd be better off buying Cointerra hardware than making their own, considering Gen 2 is at the 55nm node and probably won't be more than 4x faster per chip.

No one with exception to a private rig  http://organofcorti.blogspot.ca/2013/08/151-semi-private-mining-pool.html
has anything on the market and I'm skeptical with the estimated releases from other companies at this point.
https://ghash.io/

Your choice of cointerra is not a good sample choice it may be profitable in January but that is a long wait in bitcoin and delays are very common.
http://decentralizedhashing.com/bitcoin-mining-equipment-table/

That said I will let you choose your poison if we are going by price per G/H as you will note only AM delivers even with the current timelines of other companies any that can actually make the schedule will become interesting.

[Vycid]

It could be, but guess what? All this would make per-GH costs higher, not lower.

I've established a best-case scenario for AM and it STILL appears significantly overvalued. Thus my short position.

Cool for you, but your maths is wrong everywhere, even at $15/GH profit margin is better than 75%

You fundamentally misunderstand. That 75% is the average projected over the next year.

If Cointerra is selling at $3/GH, do you see the issue with AM producing at $15?

No, I just wanted to see how you'd react to something completely made up.

So, I show you the income statement and explain the methodology, and you continue to imply the $1.50/GH figure is "completely made up"?

Please explain your investment thesis, I'm dying to know.

Your choice of cointerra is not a good sample choice it may be profitable in January but that is a long wait in bitcoin and delays are very common.
http://decentralizedhashing.com/bitcoin-mining-equipment-table/

The fact that the stiffest competition doesn't deliver until January is the rationale behind a profit margin as high as 75%.

You must understand - all of my estimates are very conservative to offset the factors I may be missing. Yet, with those conservative estimates, the conclusion is still quite clear.

That is why I have such a high degree of certainty while everyone else appears to be purely speculating. It is extremely rare for a situation like this to appear in a market. I suspect it is a result of the bizarre wealth distribution. Many (most?) of AM's shareholders do not really understand finance.

[Vexual]

So, I show you the income statement and explain the methodology, and you continue to imply the $1.50/GH figure is "completely made up"?

Please explain your investment thesis, I'm dying to know.

Firstly, it was my statement that even at $15/GH profit margins are over 75% that was completely made up, you absorbed it like air and used it in your next conjecture.
Secondly, I've been relatively clear on how I think the data you used to arrive at $1.50/GH is wonky.
Thirdly, no.

[Vycid]

So, I show you the income statement and explain the methodology, and you continue to imply the $1.50/GH figure is "completely made up"?

Please explain your investment thesis, I'm dying to know.

Firstly, it was my statement that even at $15/GH profit margins are over 75% that was completely made up, you absorbed it like air and used it in your next conjecture.

"Used it in my next conjecture"? You mean I refuted the obviously incorrect statement "even at $15/GH profit margin is better than 75%"? I'm not sure we're reading the same message board.

Secondly, I've been relatively clear on how I think the data you used to arrive at $1.50/GH is wonky.

I can live with "wonky".

Thirdly, no.

Ah, so you're going to take pot-shots at the estimates I provide as being "wonky", but you're going to refuse to demonstrate how it'd be possible to be any more informationally efficient.

Stupendous.

[freedomno1]

If Cointerra is selling at $3/GH, do you see the issue with AM producing at $15?

Your choice of cointerra is not a good sample choice it may be profitable in January but that is a long wait in bitcoin and delays are very common.
http://decentralizedhashing.com/bitcoin-mining-equipment-table/

The fact that the stiffest competition doesn't deliver until January is the rationale behind a profit margin as high as 75%.

You must understand - all of my estimates are very conservative to offset the factors I may be missing. Yet, with those conservative estimates, the conclusion is still quite clear.

That is why I have such a high degree of certainty while everyone else appears to be purely speculating. It is extremely rare for a situation like this to appear in a market. I suspect it is a result of the bizarre wealth distribution. Many (most?) of AM's shareholders do not really understand finance.

If we assume a worst case a 30% increase in difficulty by the time the units come to market cointerra could well be seeing negative returns if difficulty rises at the speed we have been seeing recently.

This does invoke a premium for the duration of September to January and possibly longer due to the simple factor no one else has been delivering.

If Labcoin really is hashing and reveals their address by Wednesday, if Activemining and Ken start producing, if Avalon didn't delay their delivery and Klondikes were running etc.

Not even mentioning Cointerra and Hashfast or Bitfury.

Then we might say that this profit margin is ridiculous but the point of the matter is a lot of people are speaking but I have not seen any damn action in the other competitors to rile me up

I can't copy the difficulty image (Seriously long url)
But it assumes calculations for difficulty increases vs cost to buy units etc.
http://decentralizedhashing.com/bitcoin-mining-equipment-table/

Still cointerra is ranked better than a BFL Klondike and BFL never delivers anything while its still profitable my opinion on cointerra is dead bearish.

[Vycid]

If Cointerra is selling at $3/GH, do you see the issue with AM producing at $15?

Your choice of cointerra is not a good sample choice it may be profitable in January but that is a long wait in bitcoin and delays are very common.
http://decentralizedhashing.com/bitcoin-mining-equipment-table/

The fact that the stiffest competition doesn't deliver until January is the rationale behind a profit margin as high as 75%.

You must understand - all of my estimates are very conservative to offset the factors I may be missing. Yet, with those conservative estimates, the conclusion is still quite clear.

That is why I have such a high degree of certainty while everyone else appears to be purely speculating. It is extremely rare for a situation like this to appear in a market. I suspect it is a result of the bizarre wealth distribution. Many (most?) of AM's shareholders do not really understand finance.

If we assume a worst case a 30% increase in difficulty by the time the units come to market cointerra could well be seeing negative returns if difficulty rises at the speed we have been seeing recently.

Completely within the realm of possibility. I've discussed "overshoot" before. Many miners will lose money.

But if cointerra is seeing negative returns, what would you suppose that would mean for AM, considering cointerra is bringing a 28nm full custom ASIC to market?

Also, remember that cointerra's cost to produce is probably much lower than $3/GH. That's just their current sale price. There will be price wars.

Still cointerra is ranked better than a BFL Klondike and BFL never delivers anything so my opinion on cointerra is dead bearish.

... what

[freedomno1]

No, not mad. There used to be lot's of intelligent conversation here, but now it's just littered with your FUD. You might as well have your own thread.

+1

He's trying WAY too hard to be taken seriously. Ignored.

+1 for Vycid starting his own thread....

By the way networks probably passed 1petahash. Celebration time!

Celebrates thats worthy of a fb post
An Asicminer Blade is 10GH, so 1 PH is 100,000 of these.
http://thebitcoinnews.co.uk/2013/09/15/what-does-one-petahash-look-like/
http://blog.standardcrypto.com/2013/09/15/what-does-one-petahash-look-like/

Let’s imagine we are filling up One Wilshire, a famous 30 story data center in Los Angeles. Let’s say there is enough power and cooling, and we are really packing our mining equipment in.

Maybe thirty above-described rooms per story. Not every story is data center — there are lawyers and acountants who rent there too — but ignore that. So, 900 possible rooms, and 30 rooms per story. 37.8 TH/story. 1.1 PH/building.

We have our answer. The bitcoin mining network run on Gen 1 hardware consists of about one One Wilshire-sized data center filled with Blades. Somewhat confirmatory of our ballpark estimate, One Wilshire sold for 437.5 million dollars last month.

If we had this running on gen 2 hardware, one petahash would be 3 stories and the entire building would be 10 petahashes.

Either way, our building sized mine would consume 21 megawatts of power, with one third for hashing and two thirds for cooling.

And that is what one petahash (or ten petahashes of gen 2) looks like.

[freedomno1]

Still cointerra is ranked better than a BFL Klondike and BFL never delivers anything so my opinion on cointerra is dead bearish.

... what

BFL Monarch and just look at the image lol BFL Klondike  

[Vycid]

Still cointerra is ranked better than a BFL Klondike and BFL never delivers anything so my opinion on cointerra is dead bearish.

... what

BFL Monarch and just look at the image lol BFL Klondike  

No, I mean - how on earth are you making any inference about Cointerra based on the "BFL Klondike"?

[freedomno1]

Still cointerra is ranked better than a BFL Klondike and BFL never delivers anything so my opinion on cointerra is dead bearish.

... what

BFL Monarch and just look at the image lol BFL Klondike  

No, I mean - how on earth are you making any inference about Cointerra based on the "BFL Klondike"?

Both are hot air atm
Or simply we have no evidence of it working I'll leave the burden of proof on you if you can refute that

[Vycid]

Still cointerra is ranked better than a BFL Klondike and BFL never delivers anything so my opinion on cointerra is dead bearish.

... what

BFL Monarch and just look at the image lol BFL Klondike  

No, I mean - how on earth are you making any inference about Cointerra based on the "BFL Klondike"?

Both are hot air atm
Or simply we have no evidence of it working I'll leave the burden of proof on you if you can refute that

We also have no evidence of AM Gen 2 working. We assume it will happen because FC has a reputation for delivering.

Similarly, the guys at Cointerra have real industry reputations. I knew about Open Silicon before I'd ever heard of Cointerra (I'm in the semi industry) - Dr. Sherwani is the president of Open Silicon, and he is on Cointerra's advisory board. Plus, Cointerra recieved 1.5M in private equity. That typically implies they've got more than a plan to scam nerds.

These guys have almost certainly done business with GloFo before. We're up a couple quality leagues from BFL and Avalon in terms of talent.

They could be late (NB: AM could be late), but there's a buttload of competition that will fill their place if they are. Cointerra just happens to be my favorite horse in this race.

[freedomno1]

We also have no evidence of AM Gen 2 working. We assume it will happen because FC has a reputation for delivering.

Similarly, the guys at Cointerra have real industry reputations. I knew about Open Silicon before I'd ever heard of Cointerra (I'm in the semi industry) - Dr. Sherwani is the president of Open Silicon, and he is on Cointerra's advisory board. Plus, Cointerra recieved 1.5M in private equity. That typically implies they've got more than a plan to scam nerds.

These guys have almost certainly done business with GloFo before. We're up a couple quality leagues from BFL and Avalon in terms of talent.

They could be late (NB: AM could be late), but there's a buttload of competition that will fill their place if they are. Cointerra just happens to be my favorite horse in this race.

Fair point, I still am waiting on that buttload to get off their arses for now

We do not have Gen 2 proven to be hashing either so that is a valid point we are still waiting on the timeline, but I did like cointerras video and do agree with your point that they have industry reputation.
The question still remains if they will do it in time to be profitable and outpace the increases in difficulty.
http://www.youtube.com/watch?v=K199f61n4qM

As for the other buttload of competition time tells no lies but AM is still the only one that pay's  .

It will be interesting how they will add to the storyboard in the long-run
http://thegenesisblock.com/bitcoin-network-reaches-1-petahash-per-second/

2009 - CPU
2010 - GPU
2011- FPGA
2012 - Rise of The ASIC
2013 - The ASIC Arms Race
2014 - ?
Some unknown point in the future
Quantum

[Vycid]

Still cointerra is ranked better than a BFL Klondike and BFL never delivers anything so my opinion on cointerra is dead bearish.

... what

BFL Monarch and just look at the image lol BFL Klondike  

No, I mean - how on earth are you making any inference about Cointerra based on the "BFL Klondike"?

Both are hot air atm
Or simply we have no evidence of it working I'll leave the burden of proof on you if you can refute that

We also have no evidence of AM Gen 2 working. We assume it will happen because FC has a reputation for delivering.

Similarly, the guys at Cointerra have real industry reputations. I knew about Open Silicon before I'd ever heard of Cointerra (I'm in the semi industry) - Dr. Sherwani is the president of Open Silicon, and he is on Cointerra's advisory board. Plus, Cointerra recieved 1.5M in private equity. That typically implies they've got more than a plan to scam nerds.

These guys have almost certainly done business with GloFo before. We're up a couple quality leagues from BFL and Avalon in terms of talent.

They could be late (NB: AM could be late), but there's a buttload of competition that will fill their place if they are. Cointerra just happens to be my favorite horse in this race.

Fair point, I still am waiting on that buttload to get off their arses for now

We do not have Gen 2 proven to be hashing either so valid point we are still waiting on the timeline schedule for now, but personally I did like cointerras video and do agree with your point that they have industry reputation.
The question still remains if they will do it in time to be profitable and outpace the increases in difficulty.
http://www.youtube.com/watch?v=K199f61n4qM

As for the other buttload of competition time tells no lies but AM is still the only one that pay's  .

It will be interesting how they will add to the storyboard
http://thegenesisblock.com/bitcoin-network-reaches-1-petahash-per-second/

2009 - CPU
2010 - GPU
2011- FPGA
2012 - Rise of The ASIC
2013 - The ASIC Arms Race
2014 - ?
Some unknown point in the future
Quantum

The future of Bitcoin mining must be either oblivion or reversible computing. I shudder to think of the electricity wasted otherwise.

Quantum computing is many, many years off.