HashFast launches sales of the Baby Jet

[cypherdoc]

Your calculation sounds great, until you question the x365 part. Do you really think in January that you'll have another 365 days of above electricity cost hashing left? Really? You admit this yourself. I'd give you a maximum of 2 months above a Jupiter before the difficulty forces you below break even, and that's generous.

What exactly would drive difficulty that high.   People buying rigs that instantly result in a loss from day 0?

Imagine difficulty is so high a Jupiter breaks even on electrical cost ($100 in electricity = $100 in BTC).  Anything less efficient has been idled as who is going to pay $500 in electricity to get $100 in BTC.  Total network hashrate would be close to 130 PH/s (difficulty 18 billion).

Your predict is that in two months the network will double again such that a HF rig will be operating at break even electrical cost?  Really?  Other than blindly following a chart which shows an exponential growth forever have you thought about that.

a) 130 PH/s needed to double in 60 days = 2.2 PH per day in hardware.  Thats like 5,500 BabyJets or Jupiters shipping a day, every single day nonstop.

b) who would be buying these.  By your very scenario a Jupiter is no longer profitable to operate (even w/ free hardware) from day 1.  Who would be buying these hundreds of PH/s of gear to operate at a loss?  Even HF/Cointerra gear would be only marginally profitable.  Even assumming no future difficulty growth and prices as low as $2 per GH it the time to break even would be on the order of TWO YEARS.  So who is going to buy all this 130 PH/s of hardware.

c) It would take a lot more than 1 or 2 miners bad at math.  Even at $2/GH that is $2,000 per TH or $2M per PH.  Another 130 PH would be $260 million in hardware sales.  Hell pretend ASIC companies cut their prices to <$1 per GH you are still talking over $100M.

So which scenario do you think is more likely.  Difficulty continues to double forever OR at some point economics/cost starts to bend the cost curve.  Will a HF miner turn a profit?  I don't know and I am not trying to predict that.  There are far two many variables in play.  However it is painfully obvious for anyone who has done more "analysis" then clicking "calculate" on the genesis block that economic factors will eventually slow the curve.   By your logic why didn't we see difficulty grow exponentially to 100 PH/s with GPUs?  Maybe because while miners are bad at predicting the future they are pretty good at looking at the day 1 economics and that slows hardware deployment.

i fully concur and on top of that the price is starting to follow the growth of the network hashing rate in case anyone hasn't been paying attention.  there have only been only 4 days in BTC price history where the price has closed higher.

try doing these doomsday scenarios at a price of $1000/BTC.

[1713554200]

1713554200

[1713554200]

1713554200

[1713554200]

1713554200

[1713554200]

1713554200

[itod]

The equilibrium between mining income and energy costs is not "getting near."  For current gen miners, it's not even in the ballpark.  It won't be in that ballpark 'till April, *if* monthly difficulty continues to rise at 117%.  If the above holds true, you get ~1 month of extra "mining," for a net profit of pocket change.

Crumbs, I'm getting tired of explaining this, sorry but this is my last post on the subject, don't have more energy for this. End of March/April is near in my book. There comes the crucial flaw in your logic: when that moment comes, less efficient miners will be shut down, but HF/Cointerra don't get only one more month, they get *forever*. That's the key. That "forever" read very liberally, it's "forever" in mining terms meaning until more efficient chips (11nm) come to market. Please read this again and DeathAndTaxes' above post. Of course, this will not happen as strict and as mechanically as I've written, but that's general mechanism. Good day to you.

[crumbs]

...
Your predict is that in two months the network will double again such that a HF rig will be operating at break even electrical cost?  Really?  Other than blindly following a chart which shows an exponential growth forever have you thought about that.

a) 130 PH/s needed to double in 60 days = 2.2 PH per day in hardware.  Thats like 5,500 BabyJets or Jupiters shipping a day, every single day nonstop.

b) who would be buying these.  By your very scenario a Jupiter is no longer profitable to operate (even w/ free hardware) from day 1.  Who would be buying these hundreds of PH/s of gear to operate at a loss?  Even HF/Cointerra gear would be only marginally profitable.  Even assumming no future difficulty growth and prices as low as $2 per GH it the time to break even would be on the order of TWO YEARS.  So who is going to buy all this 130 PH/s of hardware.

c) It would take a lot more than 1 or 2 miners bad at math.  Even at $2/GH that is $2,000 per TH or $2M per PH.  Another 130 PH would be $260 million in hardware sales.  Hell pretend ASIC companies cut their prices to <$1 per GH you are still talking over $100M.

So which scenario do you think is more likely.  Difficulty continues to double forever OR at some point economics/cost starts to bend the cost curve.  Will a HF miner turn a profit?  I don't know and I am not trying to predict that.  There are far two many variables in play.  However it is painfully obvious for anyone who has done more "analysis" then clicking "calculate" on the genesis block that economic factors will eventually slow the curve.   By your logic why didn't we see difficulty grow exponentially to 100 PH/s with GPUs?  Maybe because while miners are bad at predicting the future they are pretty good at looking at the day 1 economics and that slows hardware deployment.

a)  Nothing needs to be shipping.  It may be ASIC manufacturers building & mining their own chips, using industrial power rates, while also shipping a fraction of that number to the same consumers who bought Asicminer gear.  The ones who like to mine at a loss to "protect the network."

b)  See above for mining by manufacturers & Asicminer buyers.

c)  Not sure.

Finally, if BTC price keeps going up, or a next-gen chip is developed, all of the calculations begin favoring continued exponential growth.
Not into indefinite future, but the timeframe we're dealing with (3-6 months).

[crumbs]

The equilibrium between mining income and energy costs is not "getting near."  For current gen miners, it's not even in the ballpark.  It won't be in that ballpark 'till April, *if* monthly difficulty continues to rise at 117%.  If the above holds true, you get ~1 month of extra "mining," for a net profit of pocket change.

Crumbs, I'm getting tired of explaining this, sorry but this is my last post on the subject, don't have more energy for this. End of March/April is near in my book. There comes the crucial flaw in your logic: when that comes, less efficient miners shut down, but HF/Cointerra don't get only one more month, they get *forever*. That's the key. That "forever" read very liberally, it's "forever" in mining terms meaning until more efficient chips (11nm) come to market. Please read this again and DeathAndTaxes' above post. Of course, this will not happen as strict and as mechanically as I've written, but that's general mechanism. Good day to you.

Oh, i get it now!  The difficulty climbs until KNC gear is unprofitable to mine, and ... doesn't climb any more.
Hashfast itself also stops filling orders & stops selling your "edge of efficiency" gear, deciding that they've made enough money.  Miner's paradise -- difficulty freeze

Of course, the other ASIC makers can't find a cheaper energy rate than the one your consumer electric co surprise butsexes you with.
Thanks for the milk

[DeathAndTaxes]

Finally, if BTC price keeps going up, or a next-gen chip is developed, all of the calculations begin favoring continued exponential growth.
Not into indefinite future, but the timeframe we're dealing with (3-6 months).

I think you are misunderstanding (giving you the benefit of the doubt that it isn't just blatant trolling)

OF COURSE difficulty is going to continue to go up for the next 3-6 months.  Nobody NOT A SINGLE PERSON said differently.  The hashrate curve will only bend when electricity and hardware costs results in a day 1 profitability that is poor.  At current prices that is >10-20 PH/s.   If prices are significantly reduced it is more like 20 to 50 PH/s.  Obviously even with rapid growth it is going to take months to get there from here.   

You make that seem like it is a bad thing.   The longer it takes the fatter the profits are.  Still you never ending refrain has been if you don't profit in the first 60 days you never will.  On inefficient hardware that might be true however if you are efficient (both in J/GH and $ per kWh) when the curve DOES bend due to economics it is possible to make a small profit every day for a very long time.   

The gold rush phase will be ending in the next couple months after that ASIC mining is going to look a lot more like GPU mining did.   There won't be front loaded profits and people will be looking at breaking even in 8-12 months.  When it is closer to 8 more hardware will be sold and difficulty will go up faster.  As it gets pushed out towards 12 people will be less interested and hardware sales will slow down.  It will become a "boring" lower margin game.

[donch]

i fully concur and on top of that the price is starting to follow the growth of the network hashing rate in case anyone hasn't been paying attention.  there have only been only 4 days in BTC price history where the price has closed higher.

try doing these doomsday scenarios at a price of $1000/BTC.

I unignored you especially :-)

There is always a short rally just before and around a difficulty change. And the price tanked just after your prediction...

P.S. The BTC market doesn't close. It's non-geographic and is open 24 hours a day!

[cypherdoc]

Still you never ending refrain has been if you don't profit in the first 60 days you never will.  On inefficient hardware that might be true however if you are efficient (both in J/GH and $ per kWh) when the curve DOES bend due to economics it is possible to make a small profit every day for a very long time.   

this is correct.  and there is really no way to model when it will occur.  as someone said above, the truth will play out somewhere in the middle of a never ending exponential ramp in hashrate and an outright crash from some negative event.

the only way to play this and still have a chance to be in the game is to buy the most efficient, cost effective hardware you can with an MPP in place to hopefully provide a cushion.

[donch]

Finally, if BTC price keeps going up, or a next-gen chip is developed, all of the calculations begin favoring continued exponential growth.
Not into indefinite future, but the timeframe we're dealing with (3-6 months).

I think you are misunderstanding (giving you the benefit of the doubt that it isn't just blatant trolling)

OF COURSE difficulty is going to continue to go up for the next 3-6 months.  Nobody NOT A SINGLE PERSON said differently.  The hashrate curve will only bend when electricity and hardware costs results in a day 1 profitability that is poor.  At current prices that is >10-20 PH/s.   If prices are significantly reduced it is more like 20 to 50 PH/s.  Obviously even with rapid growth it is going to take months to get there from here.  

You make that seem like it is a bad thing.   The longer it takes the fatter the profits are.  Still you never ending refrain has been if you don't profit in the first 60 days you never will.  On inefficient hardware that might be true however if you are efficient (both in J/GH and $ per kWh) when the curve DOES bend due to economics it is possible to make a small profit every day for a very long time.  

The gold rush phase will be ending in the next couple months after that ASIC mining is going to look a lot more like GPU mining did.   There won't be front loaded profits and people will be looking at breaking even in 8-12 months.  When it is closer to 8 more hardware will be sold and difficulty will go up faster.  As it gets pushed out towards 12 people will be less interested and hardware sales will slow down.  It will become a "boring" lower margin game.

So to boil it down to a prediction, you imagine the hashrate will stabilise around 10-20 PH/s and your Babyjet with MPP (2 TH/s) will make somewhere between $75-$150/month for a period of a year or so, beginning 90 days from Hashfast's initial delivery, assuming no one significantly beats Hashfast's power usage (via 14nm or 11nm chips)?

[itod]

Oh, i get it now!  The difficulty climbs until KNC gear is unprofitable to mine, and ... doesn't climb any more.
Hashfast itself also stops filling orders & stops selling your "edge of efficiency" gear, deciding that they've made enough money.  Miner's paradise -- difficulty freeze

Of course, the other ASIC makers can't find a cheaper energy rate than the one your consumer electric co surprise butsexes you with.
Thanks for the milk

Geographic location of the miners (regarding the electricity price) is not what was discussed. It may easily turn out to be third-world game, where electricity is dirt-cheap. Such are many "dirty" industries. HashFast may not "decide that they've made enough money", but nobody will buy miners from any company knowing machines would be unprofitable from day 1. Agian, not happening in a moment and on/off simplified like this, but some difficulty "extreme slowdown" (not freeze) is coming, when everything non-efficient is turned off.

[crumbs]

Finally, if BTC price keeps going up, or a next-gen chip is developed, all of the calculations begin favoring continued exponential growth.
Not into indefinite future, but the timeframe we're dealing with (3-6 months).

I think you are misunderstanding (giving you the benefit of the doubt that it isn't just blatant trolling)

OF COURSE difficulty is going to continue to go up for the next 3-6 months.  Nobody NOT A SINGLE PERSON said differently.  The hashrate curve will only bend when electricity and hardware costs results in a day 1 profitability that is poor.  At current prices that is >10-20 PH/s.   If prices are significantly reduced it is more like 20 to 50 PH/s.  Obviously even with rapid growth it is going to take months to get there from here.   

You make that seem like it is a bad thing.   The longer it takes the fatter the profits are.  Still you never ending refrain has been if you don't profit in the first 60 days you never will.  On inefficient hardware that might be true however if you are efficient (both in J/GH and $ per kWh) when the curve DOES bend due to economics it is possible to make a small profit every day for a very long time.   

The gold rush phase will be ending in the next couple months after that ASIC mining is going to look a lot more like GPU mining did.   There won't be front loaded profits and people will be looking at breaking even in 8-12 months.  When it is closer to 8 more hardware will be sold and difficulty will go up faster.  As it gets pushed out towards 12 people will be less interested and hardware sales will slow down.  It will become a "boring" lower margin game.

Not trolling, but finding it hard to see what, precisely, you are objecting to in the text you've quoted.  I'll break it down for clarity:

1.  If BTC exchange rate grows, mining gear remains profitable and stays online longer.  
2.  If a more efficient chip comes on the market, mining will remain profitable at a greater difficulty.

I see nothing open to debate in ether statement.  Each one is apparent on its face.
As far as boring, low-margin games go, that may or may not be the case.  I'm not arguing that.  Available data shows, on the other hand, that while mining has stopped being profitable for small miners a long time ago, most have continued to mine at a loss & continued buying gear that's virtually guaranteed to perpetuate this state.  I thank the miracle of pre-order sales, unchecked greed, & basic human inability to grasp exponential growth.

There's really not that much new here.

[cypherdoc]

Oh, i get it now!  The difficulty climbs until KNC gear is unprofitable to mine, and ... doesn't climb any more.
Hashfast itself also stops filling orders & stops selling your "edge of efficiency" gear, deciding that they've made enough money.  Miner's paradise -- difficulty freeze

Of course, the other ASIC makers can't find a cheaper energy rate than the one your consumer electric co surprise butsexes you with.
Thanks for the milk

Geographic location of the miners (regarding the electricity price) is not what was discussed. It may easily turn out to be third-world game, where electricity is dirt-cheap. Such are many "dirty" industries. HashFast may not "decide that they've made enough money", but nobody will buy miners from any company knowing machines would be unprofitable from day 1. Agian, not happening in a moment and on/off simplified like this, but some difficulty "extreme slowdown" (not freeze) is coming, when everything non-efficient is turned off.

the only logical "turn-off sequence" would place HF at the end of the line.

while Cointerra appears to have similar specs there still is no signs of a tapeout.

[crumbs]

Oh, i get it now!  The difficulty climbs until KNC gear is unprofitable to mine, and ... doesn't climb any more.
Hashfast itself also stops filling orders & stops selling your "edge of efficiency" gear, deciding that they've made enough money.  Miner's paradise -- difficulty freeze

Of course, the other ASIC makers can't find a cheaper energy rate than the one your consumer electric co surprise butsexes you with.
Thanks for the milk

Geographic location of the miners (regarding the electricity price) is not what was discussed. It may easily turn out to be third-world game, where electricity is dirt-cheap. Such are many "dirty" industries. HashFast may not "decide that they've made enough money", but nobody will buy miners from any company knowing machines would be unprofitable from day 1. Agian, not happening in a moment and on/off simplified like this, but some difficulty "extreme slowdown" (not freeze) is coming, when everything non-efficient is turned off.

Let's take this one step at a time.

"Geographic location" (industrial rates) is a part of this discussion if it factors into the hashrate endgame.  It does.  If my chips are twice as power-hungry as yours, but my power costs me half as much (including cooling), then we're neck and neck.  Industrial mines can certainly pick their location.  If you don't feel that your retail rig is competing with industrial mines for a slice of the caek, tell me why.

HashFast & co will certainly have a backlog of pre-orders stretching a few months into the past once you reach your "just efficient enough 4 mah rig to mine" sweetspot.  Those miners will come online regardless of overall profitability.  The situation today is a perfect example -- miners are expecting gear, months into the future, that they (should) realize will not break even.

Edit @ Cypher:  You're talking as if Hashfast already had their chip in hand, and the power projections have proven correct.  They do not & they haven't, respectively.

[donch]

Imagine difficulty is so high a Jupiter breaks even on electrical cost ($100 in electricity = $100 in BTC).  Anything less efficient has been idled as who is going to pay $500 in electricity to get $100 in BTC.  Total network hashrate would be close to 130 PH/s (difficulty 18 billion).

Ok, I'll admit I was mixing up my difficulty and hashrate. Easily done when dealing with so many zeroes. 139.69 is our magic number. It would be a lot more useful if the mining calculators had a hashrate column next to the difficulty one to make it easier to do a reality check.

Your predict is that in two months the network will double again such that a HF rig will be operating at break even electrical cost?  Really?  Other than blindly following a chart which shows an exponential growth forever have you thought about that.

Well my prediction was 90 days, that's when the MPP kicks in isn't it? But yes, A HF rig won't be at break even electrical cost, but neither will a Jupiter.

a) 130 PH/s needed to double in 60 days = 2.2 PH per day in hardware.  Thats like 5,500 BabyJets or Jupiters shipping a day, every single day nonstop.

b) who would be buying these.  By your very scenario a Jupiter is no longer profitable to operate (even w/ free hardware) from day 1.  Who would be buying these hundreds of PH/s of gear to operate at a loss?  Even HF/Cointerra gear would be only marginally profitable.  Even assumming no future difficulty growth and prices as low as $2 per GH it the time to break even would be on the order of TWO YEARS.  So who is going to buy all this 130 PH/s of hardware.

c) It would take a lot more than 1 or 2 miners bad at math.  Even at $2/GH that is $2,000 per TH or $2M per PH.  Another 130 PH would be $260 million in hardware sales.  Hell pretend ASIC companies cut their prices to <$1 per GH you are still talking over $100M.

So which scenario do you think is more likely.  Difficulty continues to double forever OR at some point economics/cost starts to bend the cost curve.  Will a HF miner turn a profit?  I don't know and I am not trying to predict that.  There are far two many variables in play.  However it is painfully obvious for anyone who has done more "analysis" then clicking "calculate" on the genesis block that economic factors will eventually slow the curve.   By your logic why didn't we see difficulty grow exponentially to 100 PH/s with GPUs?  Maybe because while miners are bad at predicting the future they are pretty good at looking at the day 1 economics and that slows hardware deployment.

so we get to a point where realistically, 10-20 PH/s is feasible in the next 3 months if Cointerra deliver what they promise in their business plan:

https://picostocks.com/businessplan/31

And everyone else chips in with their offerings and people still live in hope of making more than they spent.

What we end up arguing over is whether the difference between a 550GH/s Jupiter hashing from October 1st (if you were lucky) until day one of Hashfast's delivery is greater than the difference between a 550 GH/s Jupiter hashing 90 days from that day and up to 2TH/s of Babyjet's hashing at a much higher difficulty but with greater efficiency as the tail end kicks in.

It's a tricky calculation. I don't know the answer. The mining profit calculators should work backwards from a point in the future where you can fix the likely hashrate (a more logical figure to miners). The mining manufacturers should open their order books and make the market transparent...

Despite all this, I can't say for certain which unit will have made the most money in six or twelve months. I still think with this generation of ASIC, lead time in delivery has a greater effect on reaching ROI than power efficiency or delayed delivery of extra hashrate. Prove me wrong with numbers, or wait six months and show me your pool payout screenshot :-)

Maybe the next round will be different.

[flyboy]

Any updates on CHIPS or BOARDS??

or TIME LINE?

or DELAYS?

[HashFast_CL]

Any updates on CHIPS or BOARDS??

or TIME LINE?

or DELAYS?

Chips are still on schedule.   No delays so far.   Very close now!

Are we there yet?  No, but it won't be much longer.

[flyboy]

Any updates on CHIPS or BOARDS??

or TIME LINE?

or DELAYS?

Chips are still on schedule.   No delays so far.   Very close now!

Are we there yet?  No, but it won't be much longer.

So shipping still by the end of this month?

Thanks for the reply.

[DeathAndTaxes]

So to boil it down to a prediction, you imagine the hashrate will stabilise around 10-20 PH/s and your Babyjet with MPP (2 TH/s) will make somewhere between $75-$150/month for a period of a year or so, beginning 90 days from Hashfast's initial delivery, assuming no one significantly beats Hashfast's power usage (via 14nm or 11nm chips)?

No none of that was correct.   I said hashrate growth will slow and those that have high efficiency (both J/GH and $ per kWh) will be able to remain profitable for an extended period of time.   Will it be 10 PH/s? 20 PH/s? 50 PH/s? 150 PH/s?  I have no idea.  Never claimed to be able to predict the future.   

However unless you think people will continue to dump money into hardware that is DAY 0 UNPROFITABLE the difficulty growth curve will bend.  Day 0 profitability is different then difficulty rising too fast and potentially being unprofitable in the future it is more black and white and even the most delusional miner can do simple math like "at current difficulty I make $20 per day, spend $15 in electricity and thus the break even point for my $5,000 miner is 2.5 years even if difficulty doesn't rise ... so I don't think I will buy today".


BTW 14nm & 11nm are a non issue.  There won't be a 14nm SHA-256 chip until 2018 at the earliest.  There likely won't be a 22/20nm ASIC chip before 2015 and that is optimistic.   It takes roughly 3 years before a new process can even equal the cost of a prior process.   Worrying about 14nm today is just silly.

[DeathAndTaxes]

What we end up arguing over is whether the difference between a 550GH/s Jupiter hashing from October 1st (if you were lucky) until day one of Hashfast's delivery is greater than the difference between a 550 GH/s Jupiter hashing 90 days from that day and up to 2TH/s of Babyjet's hashing at a much higher difficulty but with greater efficiency as the tail end kicks in.

Why would you argue over that?  Both of those cases are sunk money already.   What matters is future sales and future profitability.   Someone who already bought from HF or KNC hopefully already did the analysis months ago and their future is already set.

[crumbs]

... However unless you think people will continue to dump money into hardware that is DAY 0 UNPROFITABLE the difficulty growth curve will bend.  Day 0 profitability is different then difficulty rising too fast and potentially being unprofitable in the future it is more black and white and even the most delusional miner can do simple math like "at current difficulty I make $20 per day, spend $15 in electricity and thus the break even point for my $5,000 miner is 2.5 years even if difficulty doesn't rise ... so I don't think I will buy today".
...

Fig. 1 depicts the difficulty at the time of purchase vs difficulty at the time same miners go online.  Please tell me how the Day 0 profitability could have been calculated.

Fig. 1

*Three months' wait is typical for mining gear sales.  KNC is not an outlier.

[donch]

What we end up arguing over is whether the difference between a 550GH/s Jupiter hashing from October 1st (if you were lucky) until day one of Hashfast's delivery is greater than the difference between a 550 GH/s Jupiter hashing 90 days from that day and up to 2TH/s of Babyjet's hashing at a much higher difficulty but with greater efficiency as the tail end kicks in.

Why would you argue over that?  Both of those cases are sunk money already.   What matters is future sales and future profitability.   Someone who already bought from HF or KNC hopefully already did the analysis months ago and their future is already set.

Well, I just continued my analysis and decided that an end-of-queue October Jupiter is pointless and cancelled the order, so if you pick a company that is refund and customer friendly, then money is not sunk. The point, which I'll repeat one final time, is that the timing of delivery is key to the speed of return and the size of return. You are dependent on all who have come become before you, and this is always a function of uncertainty with a pre-order. And queues tend to not be processed as efficiently as promised. This initial favourable window hit can far offset, delay and even nullify the "long tail" efficiency rewards you talk off. Maybe one day the "excitement" will be extracted, and the returns will be marginal and slow, but this won't happen until the premium that manufacturers are now charging has dried up.

The irony of it all is that faster and faster network speed doesn't benefit the Bitcoin network in any way. Above a certain count of nodes the network distribution is safe. This threshold was satisfactorily met when people were CPU mining, arguably more so, as it was more widely accessible. Litecoin nearly has it right, but scrypt isn't quite GPU unfriendly enough.

I'm retiring from this thread and forum section. Mining is for Emperors who need something pointed out to them about their attire :-)

Thanks for the thoughtful and considered debate and good luck!