Swedish ASIC miner company kncminer.com

[nightengale]

Do you realize that if it becomes unprofitable for Jupiters to mine, then it becomes unprofitable for anyone to mine...

This is not outside of the realm of possibility and, frankly, is my greatest concern.

[Bitcoinorama]

Question, I want to throw out there, with the benefit of hindsight, *if* KnC had followed the original path of creating a 6 gh/s (running at 7.2gh/s at the open day) Mars FPGA miner and with the BFL units and Avalon chips now having not shipped, what kind of gains would we have made?

A lot here didn't want to see that miner made, at the time it was a big risk with all that hashrate that was promised looming, but we would have had $2k off the $7k Jupiter for the Mars priced at $2,795, with Mars effectively costing us $795 in June if Jupiter launches on time...

Looking back Mars would have worked out well for all concerned, but agreeably in my mind at the time was an expensive risk to take, which I still stand by, so I don't regret my choice, but good or bad deal, presuming you could have sold off the FPGAs somehow after, maybe eBay?

[CYPER]

Do you realize that if it becomes unprofitable for Jupiters to mine, then it becomes unprofitable for anyone to mine...

This is not outside of the realm of possibility and, frankly, is my greatest concern.

Why would that be a concern?
In such a scenario where difficulty is insane (3 billion for example) then only 2 types of people will continue to mine:
people with zero running costs
people, bold enough to invest (mainly electricity) for a future profit

I can assure you that the percentage of these groups is small, compared to the ones who will just stop mining, even though they paid big bucks for hardware.

So when most of the miners stop mining, then it's not very difficult to predict what will happen - difficulty will decline by a maximum factor of 4 until it becomes sensible to mine again.

But I don't see a scenario where most miners will keep mining when profit is less than running costs If that was possible then most GPU miners will keep mining and I don't think that is the case

[nightengale]

After investing that much money in hardware, I don't see as many people dropping out as you do.

[Bitcoinorama]

Furthermore, you do realise to continuously double the hash rate, we need to continuously double the network? We are not reaching 3 billion difficulty anytime soon, it's questionable we will reach 1 billion next year.

Sure we are in for exponential hashrate increases soon, but once it settles, which it shall as new entrants in the future make proportionately less of an impact, profit will become more of a constant and predictable number. For some it really does make all the sense on the world to wait for that point, for others there are many reasons including excitement as to why they want to risk the chance to be among the early adopters of this tech. Profitable gains are in that, but not to the extent some convince themselves to believe. The risk in ROI here, is that we also have to share the brunt of NRE costs to be in that position. That's not manufacturers being greedy, that's making sure they cover their costs involved and their ass. Also if you don't want future pre-orders again they need to take into account they need to raise funds for their next gen in this lead time as well...

[WastedLTC]

I know a few of you use http://mining.thegenesisblock.com/ to estimate your return.   They had a bug in their system.  I contacted them yesterday morning and it was fixed last night.  I recommend revisiting the site.

September units still look good but October is pushing it.   Any delays after October will cause the units to not break even.  Getting the miner online the minute it's ready will be crucial.  Shipping/customs delays will be a killer.   Just to be safe, I'm signing up to host one of my units.

I can't imagine KNC shipping units that are not profitable so I really hope their 'margins on margins' work out. 

This is going to be a nail bitter for many all of us.

When you say not ROI you mean ever, or in October?!

What I was trying to point out is that the website painted a very different picture prior to the bug fix.

Using the default values for difficulty increases,  selecting the unit you purchased, and selecting the month you expect to receive.  

Selecting November returns values showing you never recouping your original investment.   The difficulty increases will probably not look like that but they may be close for 2013.   The 2014 increases don't look possible.  

[Bitcoinorama]

I know a few of you use http://mining.thegenesisblock.com/ to estimate your return.   They had a bug in their system.  I contacted them yesterday morning and it was fixed last night.  I recommend revisiting the site.

September units still look good but October is pushing it.   Any delays after October will cause the units to not break even.  Getting the miner online the minute it's ready will be crucial.  Shipping/customs delays will be a killer.   Just to be safe, I'm signing up to host one of my units.

I can't imagine KNC shipping units that are not profitable so I really hope their 'margins on margins' work out. 

This is going to be a nail bitter for many all of us.

When you say not ROI you mean ever, or in October?!

What I was trying to point out is that the website painted a very different picture prior to the bug fix.

Using the default values for difficulty increases,  selecting the unit you purchased, and selecting the month you expect to receive.  

Selecting November returns values showing you never recouping your original investment.   The difficulty increases will probably not look like that but they may be close for 2013.   The 2014 increases don't look possible.  

But that website also predicts BFL and Avalon have happy customers now, no?

[ffernandex]

The average difficulty raise since block 237888 (May 25th, 2013) per every every 2016 blocks (each difficulty adjustment period) has been 20%, which makes for about roughly 62% monthly (considering the average period for each difficulty adjustment has been 11.63 days).

Assuming this scenario maintains for the next 8 months and KnCMiner delivers in time and unit starts mining in September 30, at difficulty ~95 million, I expect one Jupiter unit will mine about 137 BTC in 6 months of operation and after that period it wont be profitable to keep paying hosting, as it will be earning around 1BTC per month and falling.

[WastedLTC]

But that website also predicts BFL and Avalon have happy customers now, no?

Lost me on that...  the only thing its taking into account is the difficulty increase based on previous months percentages, hash rate of unit, and start date.  There is nothing in the formula to account for companies releasing X hashing power on the network.

Some say that KNC being delayed just means the hashing rate spike will be delayed-- while that is true.  Bitfury will be adding huge volume along with others.. so the sooner the better (of course).

[Bitcoinorama]

The average difficulty raise since block 237888 (May 25th, 2013) per every every 2016 blocks (each difficulty adjustment period) has been 20%, which makes for about roughly 62% monthly (considering the average period for each difficulty adjustment has been 11.63 days).

Assuming this scenario maintains for the next 8 months and KnCMiner delivers in time and unit starts mining in September 30, at difficulty ~95 million, I expect one Jupiter unit will mine about 137 BTC in 6 months of operation and after that period it wont be profitable to keep paying hosting, as it will be earning around 1BTC per month and falling.

Unless by some unbelievable coincidence, the public start to embrace Bitcoin through easier access to exchange, useful or know sites adopting it as a payment method, or some corrupt central bank shenanigans pushing BTC price upto a point where it once becomes profitable.

There's also the fact energy companies could drop prices, but haha that's NEVER gonna happen, not in the UK at least, where they caused outrage last winter with privatised energy companies profiteering and forcing pensioners to either starve or freeze. Ironically it may well be worth donating your old units to them so at least they get free heating!

[Bitcoinorama]

But that website also predicts BFL and Avalon have happy customers now, no?

Lost me on that...  the only thing its taking into account is the difficulty increase based on previous months percentages, hash rate of unit, and start date.  There is nothing in the formula to account for companies releasing X hashing power on the network.

Some say that KNC being delayed just means the hashing rate spike will be delayed-- while that is true.  Bitfury will be adding huge volume along with others.. so the sooner the better (of course).

Sorry assumed this calc worked by factoring in release dates of new tech, thus Avalon's bulk chips and BFL's anything was delivered by now...

[kano]

...It's a PCB board and some chips...

Then if everything is on track, when can we realistically expect to see a prototype board and chips hashing? Unless KnC is running significantly behind, or even ahead of schedule, this is not an unreasonable request...

Sometime in September, this was said at the open day.

The design process we were told was to make equivalent of the final device but plug in some low hashing FPGA boards that they'd send us and then we (KnC, ckolivas and myself) would be able to work on ironing out any kinks before the ASIC chips turned up.
Of course with the ASIC chips there would no doubt be more 'ironing' but most likely minimal.
That was KnC's idea.
We're still waiting to find out when that might happen ...

[CYPER]

The average difficulty raise since block 237888 (May 25th, 2013) per every every 2016 blocks (each difficulty adjustment period) has been 20%, which makes for about roughly 62% monthly (considering the average period for each difficulty adjustment has been 11.63 days).

20% each difficulty adjustment period is not the same as 62% monthly
You don't just divide the number of days by the difficulty adjustment period in days and multiply by 20%

If you have 20% increase each 11.63 days then in 116 days that won't amount to 200% total increase, but much more

[crumbs]

Furthermore, you do realise to continuously double the hash rate, we need to continuously double the network? We are not reaching 3 billion difficulty anytime soon, it's questionable we will reach 1 billion next year.

If by "double the network," you mean "double the hashrate," then yeah.  If you mean "double the number of nodes," then no.  
Continuously doubling hashrate is not just plausible, it's likely.  If just a fraction of the chipper's claims turn out to be true, the hashrate is going to do that & more for quite a while.  Keep in mind that we're talking ASICs -- once the R&D is done, and production ramped up, the per-chip cost is nominal, as is required support circuitry.  Consumer prices will reflect total profits, not per-unit profits.  If by selling your chips at 1/100th the price you can sell x101 more chips, that's what you do.

If you have competition, you may be forced to sell your chips at a loss, and a substantial loss, or not be able to sell any *at all.*  
As far as miners shutting down & waiting for difficulty to drop, as someone mentioned above?  Not even going there.

TL;DR:  Doubling hashrate?  Not just possible, almost assured.

[-Redacted-]

The average difficulty raise since block 237888 (May 25th, 2013) per every every 2016 blocks (each difficulty adjustment period) has been 20%, which makes for about roughly 62% monthly (considering the average period for each difficulty adjustment has been 11.63 days).

20% each difficulty adjustment period is not the same as 62% monthly
You don't just divide the number of days by the difficulty adjustment period in days and multiply by 20%

If you have 20% increase each 11.63 days then in 116 days that won't amount to 200% total increase, but much more

About 600% unless I mis-factored somewhere along the way.  That "compound interest" is a real bitch...

[maxpower]

The average difficulty raise since block 237888 (May 25th, 2013) per every every 2016 blocks (each difficulty adjustment period) has been 20%, which makes for about roughly 62% monthly (considering the average period for each difficulty adjustment has been 11.63 days).

Assuming this scenario maintains for the next 8 months and KnCMiner delivers in time and unit starts mining in September 30, at difficulty ~95 million, I expect one Jupiter unit will mine about 137 BTC in 6 months of operation and after that period it wont be profitable to keep paying hosting, as it will be earning around 1BTC per month and falling.

Unless by some unbelievable coincidence, the public start to embrace Bitcoin through easier access to exchange, useful or know sites adopting it as a payment method, or some corrupt central bank shenanigans pushing BTC price upto a point where it once becomes profitable.

There's also the fact energy companies could drop prices, but haha that's NEVER gonna happen, not in the UK at least, where they caused outrage last winter with privatised energy companies profiteering and forcing pensioners to either starve or freeze. Ironically it may well be worth donating your old units to them so at least they get free heating!

I wonder whether markets could become more stable for some of the SHA256 altcoins when BTC difficulty gets prohibitively high. There are already periods of time (according to coinchoose.com, at least) where PPCoin and TerraCoin are more profitable to mine than Bitcoin--but because there's so much variance in difficulty and exchange rates, it's hard to predict right now. But what if we see a scenario where smaller miners start to point their ASICs at altcoins? Maybe that would improve stability of those markets and extend the lifetime of the early hardware a bit.

I'd be curious to know whether miners with small-to-midsize GPU or FPGA farms are already starting to do this.

[CYPER]

The average difficulty raise since block 237888 (May 25th, 2013) per every every 2016 blocks (each difficulty adjustment period) has been 20%, which makes for about roughly 62% monthly (considering the average period for each difficulty adjustment has been 11.63 days).

20% each difficulty adjustment period is not the same as 62% monthly
You don't just divide the number of days by the difficulty adjustment period in days and multiply by 20%

If you have 20% increase each 11.63 days then in 116 days that won't amount to 200% total increase, but much more

About 600% unless I mis-factored somewhere along the way...

I've just ran some calculations and between 25th May 2013 and 5th August 2013 difficulty has increased 208.62%, so he might be correct with the 62% monthly figure.

[CYPER]

But what if we see a scenario where smaller miners start to point their ASICs at altcoins?

Not many SHA-256 alt coins to choose from and the most profitable compared to Bitcoin are scrypt, so it makes no sense to point ASIC to SHA-256 alt coins.

[HyperMega]

On a serious note though; what's the latest on KNC's progress?
Does the October shipping date seem likely?

We have no reason to believe they won't hit their ship-date at this point. The only evidence we have that they will is their claim that they are on time with everything so far.

Yes that is true, but for that kind of project they are really quiet about their project schedule. (Yes, I know: "The competitors!" )

You seem pretty smart.   Can you tell me what possibly a competitor could do by knowing that you have finished your tape out?   How would that help them do anything?   Or that you have populated your board and are testing?   What would that information do for a competitor?   Would it help them get their design done faster?   Would it help them test their design?   Would it help them move up in the line at the foundry?
"competitors":  Everytime I read that I think "LIARS".  "BULLSHITTERS"

No, to be honest I have absolutely no idea how announcing the completion of concrete milestones including dates should help any competitor, when you are on track for September delivery.

They must have a internal project plan with dates for all the major milestones (there are a lot of them), which finally ends with delivery to customers end of September, which is not that far away.

From my point of view, being more transparent here could only be beneficial for KnC with respect to their reputation and a kind of "shock and awe" for any competitor.
Some of them would maybe stop development completely, because they will realize that they are too late in game and that nobody will buy their products when KnC already has 28nm chips on the market for several months. But this would only be valid, if you have an realistic plan and if you are 100% on track.

[Bitcoinorama]

But what if we see a scenario where smaller miners start to point their ASICs at altcoins?

Not many SHA-256 alt coins to choose from and the most profitable compared to Bitcoin are scrypt, so it makes no sense to point ASIC to SHA-256 alt coins.

But you can merge mine, and there's a poll on the KnC homepage currently if you wish them to consider that. It was added yesterday.