Realistic difficulty Projections

[Hqen2000]

Realistically it will come all come down to power consumption and difficulty (assuming you already own a miner). If you look at the 'calculators' in a year my Antminer S1 will make $2 a month in BTC and consume $22/month in electricity. Clearly nobody is going to mine at a tremendous loss, in fact at the current projection rates it will no longer be profitable to run an Antminer S1 in about a month from now (depending on your individual variables it may not be profitable now).

Will the difficulty really increase exponentially at a rate of 50% a month? I seriously doubt that. Eventually more people will bleed off, and only those who have very efficient setups, cool temps, and cheap power will be able to profit from mining.

For those who live in areas where it is not feasible to mine long-term, you may want to consider selling your equipment to someone who can while the difficult rate is still reasonable and you can still sell your rig for a decent amount.

So are we likely to see the hash rate all much from where it is now? Difficult increases? Where do you think it will plateau?

[bitco.in]

I thought difficulty increases were relatively predictable already.

[Hqen2000]

My understanding is that difficulty is based on the current block solution rate, and it is regulated to be around 10 minutes. Right now it is 7-8 minutes, so the difficulty will have a ~20% increase to compensate. But if the hash rate was lower it does not need to increase by 20% it can be anything. The difficulty can even decrease.

Again that's my understanding, i'm relatively new so if I'm wrong someone please correct me.

[Tron]

What tends to happen is that money flows out of Bitcoin into mining gear which causes the difficulty to rise as that mining gear comes online.  It also pushes the value per Bitcoin down because the mining gear companies do not hold, but sell to convert the value into labor, material, shipping, etc.  At some point the difficulty is so high and the value of a Bitcoin so low that it becomes unprofitable or barely profitable to mine.   Some miners drop out.

At that point it makes more sense to buy Bitcoin because they are very 'expensive' to mine in terms of resources (capital equipment and electricity).  The price of Bitcoin then goes up, which makes mining profitable again.

This balanced feedback loop is by design, and is just one of the brilliant breakthroughs of the Bitcoin protocol.    As this wash-rinse-and-repeat loop continues, the security and reliability of the network increases.

Here is something to consider if you are contemplating mining.  IF/WHEN the price goes up again, there will be a scramble for mining equipment as it becomes profitable again.  If you already have it, you can take advantage.  If not, you'll get in line for pre-orders as everyone else also tries to get mining equipment.   Once everyone gets their equipment and starts mining, the difficulty will jump and the cycle will begin again.

In the short term you can do really well by having mining gear in-place.  In the long run, it is the relative advantages that will win out.  What is your cost of electricity for mining and cooling relative to the rest of the mining community?  What is your GHash/Watt relative to the rest of the miners?  Are there other factors such as the need to heat your home?

Tron Black
CEO, CoinCPA.com

[Syke]

in fact at the current projection rates it will no longer be profitable to run an Antminer S1 in about a month from now (depending on your individual variables it may not be profitable now).

Let's go with 180 GH/s at  6 B difficulty and $450 for a starting point and $.15/kwh.

Income: $6.75
Expense: $1.30

S1 is still quite profitable for everyone right now.

Go out several months at 20 B difficulty.

Income: $2.00

Still profitable for nearly everyone. A lot of miners are paying less than $.15.

Difficulty isn't that hard to predict. It's going to keep growing at 10%-20% per adjustment for several months. What's quite unpredictable is the exchange rate. It could drop in half or double in mere days. If summer hits and the exchange rate keeps dropping, then a lot of miners are going to hit that unprofitable period. Like you said, miners with high electrical costs are going to need to sell their hardware to miners with low electricity. On the other hand, there's a very good chance of the exchange rate jumping up in a few months, which will keep everyone profitable.

[thefunkybits]

The network's current growth rate is impossible to keep up. Unless BTC hits $10,000 this year (and even if it does) the network is not going to scale as much as it has. Im gonna predict a 10,000,000,000 difficulty by June

[Hqen2000]

looking historically it's only bee this year, that the hash rate has tripled in size, much of that probably had to do with the price increases. 10,000,000,000 by june is probably pretty realistic.

One of the only ways to increase the hash rate is to more devices to enter the network. Which means people would have to buy a new miner (not one that someone else was mining with). Who would buy a S2 type system that will make $30 a month and cost them $50 on in power, even if the got the system really cheap.

I'm just not seeing the huge increases in hash rate anymore. The newer systems are more efficient but also a lot more expensive. I think once everyone sees what it is like to mine at 10,000,000,000 huge power bills very little profit the hash rates may begin to fall. Of course depending on the cost of the coin at that time.




Mar 24 2014   5,006,860,589   17.80%   35,840,504 GH/s
Mar 13 2014   4,250,217,920   11.39%   30,424,245 GH/s
Feb 28 2014   3,815,723,799   21.92%   27,314,015 GH/s
Feb 17 2014   3,129,573,175   19.39%   22,402,357 GH/s
Feb 05 2014   2,621,404,453   19.49%   18,764,744 GH/s
Jan 24 2014   2,193,847,870   22.59%   15,704,175 GH/s
Jan 13 2014   1,789,546,951   26.16%   12,810,076 GH/s
Jan 02 2014   1,418,481,395   20.12%   10,153,885 GH/s
Dec 21 2013   1,180,923,195   30.01%   8,453,378 GH/s
Dec 10 2013   908,350,862   28.41%   6,502,229 GH/s
Nov 29 2013   707,408,283   16.07%   5,063,826 GH/s
Nov 17 2013   609,482,680   19.29%   4,362,847 GH/s
Nov 05 2013   510,929,738   30.70%   3,657,378 GH/s
Oct 26 2013   390,928,788   46.02%   2,798,377 GH/s
Oct 16 2013   267,731,249   41.45%   1,916,495 GH/s
Oct 06 2013   189,281,249   27.19%   1,354,928 GH/s
Sep 25 2013   148,819,200   32.13%   1,065,289 GH/s
Sep 14 2013   112,628,549   29.56%   806,227 GH/s
Sep 04 2013   86,933,018   32.22%   622,291 GH/s
Aug 24 2013   65,750,060   29.40%   470,657 GH/s
Aug 13 2013   50,810,339   35.88%   363,715 GH/s
Aug 03 2013   37,392,766   19.63%   267,668 GH/s
Jul 22 2013   31,256,961   19.47%   223,746 GH/s
Jul 11 2013   26,162,876   22.63%   187,281 GH/s
Jun 29 2013   21,335,329   10.32%   152,724 GH/s
Jun 16 2013   19,339,258   23.92%   138,436 GH/s
Jun 05 2013   15,605,633   28.41%   111,709 GH/s
May 25 2013   12,153,412   8.64%   86,998 GH/s
May 12 2013   11,187,257   11.03%   80,082 GH/s
Apr 30 2013   10,076,293   12.28%   72,129 GH/s
Apr 17 2013   8,974,296   16.96%   64,241 GH/s
Apr 05 2013   7,673,000   14.59%   54,925 GH/s
Mar 24 2013   6,695,826   38.13%   47,931 GH/s
Mar 14 2013   4,847,647   10.98%   34,701 GH/s
Mar 01 2013   4,367,876   19.63%   31,266 GH/s
Feb 18 2013   3,651,012   11.47%   26,135 GH/s
Feb 05 2013   3,275,465   10.33%   23,447 GH/s
Jan 23 2013   2,968,775   -8.64%   21,251 GH/s
Jan 08 2013   3,249,550   9.06%   23,261 GH/s
Dec 26 2012   2,979,637   -11.59%   21,329 GH/s
Dec 10 2012   3,370,182   -2.00%   24,125 GH/s
Nov 26 2012   3,438,909   2.08%   24,617 GH/s
Nov 12 2012   3,368,767   1.95%   24,115 GH/s
Oct 30 2012   3,304,356   7.55%   23,654 GH/s
Oct 16 2012   3,072,322   0.58%   21,993 GH/s
Oct 03 2012   3,054,628   6.65%   21,866 GH/s
Sep 19 2012   2,864,141   6.31%   20,502 GH/s
Sep 06 2012   2,694,048   10.38%   19,285 GH/s
Aug 25 2012   2,440,643   11.40%   17,471 GH/s

[Syke]

The network's current growth rate is impossible to keep up. Unless BTC hits $10,000 this year (and even if it does) the network is not going to scale as much as it has. Im gonna predict a 10,000,000,000 difficulty by June

Yup, probably in that ballpark. But that's not the top. It's going to keep going, and going, and going.

[davejh]

A difficulty of 10B by June seems almost certain. We're about to see a > 20% jump in the difficulty again in the next 24 hours. Realistically to hit 10B we need a total hashing rate of 71 PHash/s and we're already peaking at > 50 in the last few days.

There are still plenty of companies scheduled to deliver 28 nm ASICs in the timeframe.

I did a quick calculation on the possible hash rate if everyone was able to convert current returns into hardware and got a number around 120 PHash/s and that didn't account for any new designs that are more cost effective. A more realistic number is higher and it's higher again if anyone is stealthily building ASICs to use for mining themselves (their cost base will be much lower).

[Operatr]

Daily growth rate is slowing down, from the high of 4% with the influx of ASICs now down to 1.5% and falling.

The miner base is actually relatively small, and at some point we will reach a saturation. All of the miners sold are not necessarily adding to the overall hashrate, as they may just be replacing older miners with more efficient ones, or mining on other networks entirely. ASIC suppliers also have a certain cap as far as how much they can supply, so network growth is also limited by this too at some point. Power usage is also a factor, as dense farms eat a lot of juice that at a point becomes very expensive to implement. Most data centers are not built to handle that kind of power load at upwards of 1000w per miner for the industrial ones. KnC and others are building facilities from the ground up at significant expense.

28nm is about as low as ASIC development can go as well, so the power/efficiency increases are effectively capped where they are for now. 20nm does exist which is mostly CPU production, but these chips are very difficult and expensive to make. I think we'll have 28-40 nm chips for a long time to come with so much already invested into them. If a supplier did create 20nm chips right now they would be insanely expensive and wouldn't offer that big of an advantage over 28nm. Between AsicMiner Gen 1 and the latest 28nm chips there is a big difference, but now those differences are getting smaller and smaller between generations.

Realistically hashrate can't rise this fast forever, and does in fact seem to be slowing down now. Though, in general with Bitcoin, expect the unexpected.

[samsonn25]

in fact at the current projection rates it will no longer be profitable to run an Antminer S1 in about a month from now (depending on your individual variables it may not be profitable now).

Let's go with 180 GH/s at  6 B difficulty and $450 for a starting point and $.15/kwh.

Income: $6.75
Expense: $1.30

S1 is still quite profitable for everyone right now.

Go out several months at 20 B difficulty.

Income: $2.00

Still profitable for nearly everyone. A lot of miners are paying less than $.15.

Difficulty isn't that hard to predict. It's going to keep growing at 10%-20% per adjustment for several months. What's quite unpredictable is the exchange rate. It could drop in half or double in mere days. If summer hits and the exchange rate keeps dropping, then a lot of miners are going to hit that unprofitable period. Like you said, miners with high electrical costs are going to need to sell their hardware to miners with low electricity. On the other hand, there's a very good chance of the exchange rate jumping up in a few months, which will keep everyone profitable.

You are forgetting the initial cost of the machine $450.  This unit will make less than $250 btc at current prices in its lifetime. By mid july will cost more to run than what it mines.

[Crossbow376]

in fact at the current projection rates it will no longer be profitable to run an Antminer S1 in about a month from now (depending on your individual variables it may not be profitable now).

Let's go with 180 GH/s at  6 B difficulty and $450 for a starting point and $.15/kwh.

Income: $6.75
Expense: $1.30

S1 is still quite profitable for everyone right now.

Go out several months at 20 B difficulty.

Income: $2.00

Still profitable for nearly everyone. A lot of miners are paying less than $.15.

Difficulty isn't that hard to predict. It's going to keep growing at 10%-20% per adjustment for several months. What's quite unpredictable is the exchange rate. It could drop in half or double in mere days. If summer hits and the exchange rate keeps dropping, then a lot of miners are going to hit that unprofitable period. Like you said, miners with high electrical costs are going to need to sell their hardware to miners with low electricity. On the other hand, there's a very good chance of the exchange rate jumping up in a few months, which will keep everyone profitable.

You are forgetting the initial cost of the machine $450.  This unit will make less than $250 btc at current prices in its lifetime. By mid july will cost more to run than what it mines.

It is profitable to run the ASIC.
It is not profitable to buy the ASIC.

[samsonn25]

Yes.   

But how do we acquire the machine without having to spend money to purchase it.

[Crossbow376]

Yes.  

But how do we acquire the machine without having to spend money to purchase it.

There is no way.

So, if you already bought an ASIC, it is still reasonable to keep running it for quite some time, and you can sell it to newbies on eBay in the future.
If you haven't bought an ASIC yet, you should calculate the profitability again and you shouldn't buy the ASIC at all.

[Hqen2000]

What should you do instead? Just not mine?

[davejh]

Daily growth rate is slowing down, from the high of 4% with the influx of ASICs now down to 1.5% and falling.

The miner base is actually relatively small, and at some point we will reach a saturation. All of the miners sold are not necessarily adding to the overall hashrate, as they may just be replacing older miners with more efficient ones, or mining on other networks entirely. ASIC suppliers also have a certain cap as far as how much they can supply, so network growth is also limited by this too at some point. Power usage is also a factor, as dense farms eat a lot of juice that at a point becomes very expensive to implement. Most data centers are not built to handle that kind of power load at upwards of 1000w per miner for the industrial ones. KnC and others are building facilities from the ground up at significant expense.

28nm is about as low as ASIC development can go as well, so the power/efficiency increases are effectively capped where they are for now. 20nm does exist which is mostly CPU production, but these chips are very difficult and expensive to make. I think we'll have 28-40 nm chips for a long time to come with so much already invested into them. If a supplier did create 20nm chips right now they would be insanely expensive and wouldn't offer that big of an advantage over 28nm. Between AsicMiner Gen 1 and the latest 28nm chips there is a big difference, but now those differences are getting smaller and smaller between generations.

Realistically hashrate can't rise this fast forever, and does in fact seem to be slowing down now. Though, in general with Bitcoin, expect the unexpected.

20 nm will start to get more interesting later this year although it's not clear how much fab capacity Bitcoin ASICs will get. 20 nm gives some scope for almost a 2x improvement over 28 nm and at the moment I suspect no-one has really pushed the envelope on power-efficient designs yet so I wouldn't be surprised to see 2x to 3x there too. Once we get slowing hashing rate increases then I'm pretty sure power efficiency will become the next big battleground so I'd guess we'll still see innovation there even in 28 nm.

I'm not sure that there's a real cap on the ASIC vendors other than the amount of money flowing to the miners who can then spend large amounts of it on hardware; changing that really takes a BTC price increase though as transaction fees just don't add much to the pot. Once new generations of hardware can't just get a 10x win from process technology changes then the obvious thing for these companies to do is keep producing the same ASICs for longer. They've already paid all of the NREs and fab costs always start to fall as a process matures so they can afford to sell for longer and at reduced prices while still retaining margins - in fact they'll pretty much have to do that to keep revenue coming in.

My guess is that current designs and mining rewards can very quickly get us to somewhere in the range of 100 to 150 PH/s without large-scale power-efficient mining and then we'll see a much more incremental approach. If increases reduce to around 10x per year though then it starts to become possible to look at an ROI over a much longer timescale and where depreciation of hardware is over a much longer timescale and then data-center-scale mining will really start to make sense.

[Puppet]

only 10B in June? Hashrate levelling off? I say keep dreaming.

Even if no one orders anything new anymore, the amount of placed but undelivered orders alone is already staggering. BFL hasnt shipped a single 28nm product yet,  ditto for BlackArrow and Spondoolies. Then you have the companies that still have to work through months of backlogs and promised upgrades, like Bitmine, Cointerra, HF,..

Last but not least, there is Asicminer:

1) What is the order of magnitude of 3rd generation chips ASICMiner anticipates to sell (total / first batch / first 3 months)?
re 1) Order of magnitude: Depending on fab capacity. Total target is 10k to 40k wafers, relying on competitors, Bitcoin price, and other factors. A wafer is rated at 40 TH/s.

https://bitcointalk.org/index.php?topic=99497.msg5025133#msg5025133

For the math impaired 10K wafers x 40TH=400PH or almost 10x todays overall network hashrate. Leveling off, my ass.

[samsonn25]

There are millions and millions  of 40nm 55 nm  56nm and 65nm  chips out there too now that noone talks about.