Topic: Bitmain's Released Antminer S9, World's First 16nm Miner Ready to Order - page 39. (Read 531164 times)

New batch just was announced and it's still $2320, Bitmain is crazy with these prices when BTC is so low.

Hopefully Samsung wakes up their idea and realizes that there is no point competing with bitmain in such a small market.... No one wins then.

I think so too,
I have a suspicion that the batches they release to the public are just the leftovers...

Yikes, that's a lot of chips for sure. Thanks for that link. I was estimating based upon the number of miners hitting the market we see from here.

They must be supplying their own mining operations and partner centers, it doesn't look like the S9 sales from their website are many at all.

While I agree with everything else in your post, the statement above is troubling me. It has to be a lot more than 7000 units in 60 days. Nvidia and AMD for example both produce millions of GPU's each month.

Lets do some research:
Bitmain's S9 production numbers are not truly known but we can make some educated guesses, we know for example that Bitmain is currently buying 100,000 ASIC chip wafers from TSMC monthly (source: https://www.digitimes.com/news/a20180130PD208.html?mod=3&q=TSMC).
You get roughly 5158 chips out of one wafer so at 189 chips per S9, that's 27 S9's per wafer (source: https://medium.com/@jimmysong/just-how-profitable-is-bitmain-a9df82c761a)
This means that with the 100,000 wafers they order each month, they can make 2.7 million S9's monthly!

I'm sure not all those chips/wafers are destined for S9's but it makes the 857,000 S9's over 60 days you calculated above sound much more plausible.

Dude, it is foolish to assume that btc will go back to 15k usd. There is no guarantee of that.... Past performance is not an indication of future performance.

That is very possible. Mining is  not good now. Alot of things can happen before breakeven occurs. Currently, break even is longer and thus, risks are far higher.

No point in getting too excited. Those numbers are highly conservative and simply to pose as an example to the newbies that are unaware that selling Bitcoin at 7500 USD/ BTC is foolish.

When I first started this project and keeping the spreadsheet going we were having 30% to 20% gains on average every difficulty target.

What I wanted to demonstrate with this chart that even if the network balanced out and remained at an improbable 1% gain. At the present rate 7500 USD you'd be hard-pressed even make one Bitcoin profitably. I have also set one to a more reasonable 10% increase for the next 6 retargets while BTC is trading at 15000 usd

I also make note that anything past 3 difficulty retarget is hard to predict. For reference I set a second difficulty at 1% after 6 target levels.. That is why I try to take the average of 2016-2017 and the last month of 2018.

I personally believe that come April we're going to see another 20% increase. This is simply because Samsung is now making a ASICs. With increased production we are also due for an increase in chip performance with the next-generation hardware.

Update: For instance if you mined with the S9 under those conservative network difficulty increases you would net 0.49 BTC and at 15000 = 4850 USD.
After hardware and power costs your profit would 2500usd.

This doesn't even factor in overhead such as employee security and rent or any incidentals.

If you took that same investment and purchased BTC at 7500usd now. You would have 0.65btc.

So ultimately selling BTC now for hardware seems foolish. That is unless instead of paying tax you prefer to have that tax portion work for you by reinvested in a mining asset you can depreciate.

I am well capable of reading, here’s what was said:

Have you had a look at the curve over the last 2 years: https://data.bitcoinity.org/bitcoin/difficulty/2y?t=l
As HoleShot correctly points out in the previous post, it’s following a logarithmic function.
What you are doing is on this graph, you’re drawing a straight line between the first and last point, you’re then extending that line to predict the future. Surely you can see that the slope of the line you’re drawing is not following the latest trend of the curve.

Perhaps this one: https://blockchain.info/charts/hash-rate
or this one: https://bitinfocharts.com/comparison/bitcoin-hashrate.html#1y
or this one: https://www.coinwarz.com/network-hashrate-charts/bitcoin-network-hashrate-chart,
or this one: https://data.bitcoinity.org/bitcoin/hashrate/6m?c=m&g=15&t=a
…
Which one are you looking at?

Good.

Not sure why you would expect that, growth rate has always increased.

Very wise decision Sir, instant 12k profit :-)

Learn to read, 5% preset difficulty every 14 days, thats an avg of 11% monthly, the avg was 7% in 2017. Last year there was some big jumps in difficulty too...

And like i told you some posts before i dont know what hashrate chart u are looking at. The one im looking at we are back at hashrate lvls of 15 days ago and it keeps going down.

Also nobody expects difficulty to simply go flat but to go up in lower numbers than it is raising after the whole craze in december

PD: I dont think anybody is discussing with you that selling a s9 for 5k is much better than mining i sold my order of 6 for 4.5k each

It's actually a logarithmic function relative to the total hash power on the network. Difficulty actually chases the total power around in response. That's should be but alas it's not good news.

Difficulty will double in response to the total power doubling which will result in you getting halve the reward per hash. You can tell when the difficulty will increase anytime you see the time between blocks below 10 minutes. Since it is now at 8.23 minutes difficulty must increase. Capacity is increasing so quickly that the time between blocks is not changing quickly enough to accurately regulate the rate so it's been below 10 minutes for awhile.

I first thought it would take a long time for the network power to double. I just checked an OMG it doubled in about 60 days.

That is roughly the equivalent of 857,000 S9 miners. I'd be surprised if Bitmain shipped 7,000 miners in the last 60 days. It is obvious that the number of enterprise mining facilities going on-line is epic. The cost and availability of S9s is completely irrelevant to the difficulty.

I read a story where a Canadian Hydro-Electric power company conceded they will not be able to service all the mining applications they received. Holy electron flow Batman. Some of that was from Chinese miners looking to relocate but not all. They said they are getting multiple applications per day, none of which can they supply. Other power companies are starting to deny applications for political reasons.

The law of large numbers, power generation and politics will eventually lengthen the time it takes to double capacity. If the same number of facilities go online in the next 60 days, that'll only be a 50% increase to the total network.

Since the peak in December net profits from S9s was cut in 1/2 by network capacity and 1/3 by BTC value and a bit by transaction fees reduction.

The only way to save the S9 profitability is to increase the value of BTC, or anything else SHA-256 mine-able. If BTC value goes to $100k we'll all be feeling smart again.




 

On the contrary sir, I understand perfectly. That's what started this whole thing, just look at my first posts in this thread.

As I mentioned many times before in this thread, I was mining with my beloved S9 till just over a month ago.
I saw they were selling on Ebay for ridicules prices and did some calculations to see if my miner would still make me as much money as the profit I was able to sell it for at that time. I posted the resulting spreadsheet here, the outcome was obvious, SELL!
So after mining for a while and making some descent BTC, I ended up selling my 6 month old miner for $2600 more than I originally bought it. I'm now investing that profit directly into crypto (not BTC at this time but that's another discussion).

Anyway, since then, miners on Ebay have halved in price. I sold at the peak it seems.

So the question for you is, do you sell your miner(s) today and probably be able to get your full original investment back (that's what they're selling for roughly now on Ebay) on top of the BTC you earned so far, or,
do you wait a few months and sell it/them for perhaps a couple of thousands less? And how much BTC will your miner mine in the meantime to make up for that difference.
It's a hard decision in any case, mining is just fun, I know! Yet, rationality should prevail.

I had not even seen that last 10% increase, we can add it to the list :-)

Current difficulty:    2,874,674,234,416
January 26 difficulty:    2,603,077,300,218

+10.4337 %

I got your point from the very beginning. Unfortunately many cant.

A 1% increase every 18 days, you guys are kidding, right?

2 weeks ago it increased by 16.8%,
2 weeks before that it increased by 15.3%,
2 weeks before that it increased by 3.0%,
2 weeks before that it increased by 17.7%
2 weeks before that it increased by 18.1%

In the last year alone, difficulty has increased by 490% (last Feb it was 0.441T, now it’s 2.6T), and it is now increasing at a rate much faster than that.
So I don’t know how you guys justify the numbers above.

And what’s this about “difficulty will level out”? Why would you think that?
You are aware that in the almost 10 years of bitcoin history, THAT HAS NEVER HAPPENED!
In fact, apart from the odd blip here and there, difficulty has ALWAYS increased and the rate at which it has been growing has ALWAYS been increasing as well.

Just look at the data:
https://data.bitcoinity.org/bitcoin/difficulty/5y?t=l

Saying it will magically start to flatten out is true wishful thinking in my opinion.
Anyway, keep buying those S9’s every time Bitmain releases a batch…

I do apologize as I had posted the sheet ahead of completing the difficulty estimates. However it still may prove better to buy hardware as the difficulty will level out as adding new hardware to the network does subside.



Here is an example from the above linked document with the new offering from Bitmain.

^ UPDATE: Now with two difficulty dates. Presently set at 5% and 1% on the 6th month.

2ND UPDATE: I added a result at 15,000.00 USD to BTC Estimated Return with a Network difficulty increase of 10%

Great work, goes to show some naysayers here talking about Avg 15% increases... Some ppl are very naive to think big mining operations will run at a loss based on faith of a bitcoin price increase. Fact is mining will always be profitable, the only way a S9 becomes unprofitable is if something REALLY good is released in the next few months and they flood the market with it.

Safe is relative to how much ASIC hardware will hit the network by then. Additionally if your objective is to accumulate more BTC then it seems like a small risk at this point. I also factored the max difficulty before it would become unprofitable.

However it seems like a good offering.

Also if you liked the work then +1

Amazing research man, great job! Can't wait to see the new calculations with the yet-to-be-released chips.
Anyway, if I am reading this right, ROI of 1 S9 will be 1 year (if the current BTC price remains unchanged) by November 2018.

It is more than highly unlikely to have the same BTC price by Nov, so I think that all S9 owners are pretty safe, right?

Here is a chart that you can use that will help.

I have updated it with Bitmains latest offering of the Antminer S9.


It seems that the network difficulty will be dropping. Perhaps due to BCH taking a little of the mining hardware out of the picture or older hardware being taken offline to to operating costs.

To clarify the chart I have is set to have an increase in difficulty at 1% every 18 days. THIS IS AN ESTIMATE. It may drop % a few times. However I have factored in the average increase over the last two years and even the average from the last year.

https://docs.google.com/spreadsheets/d/11QS1BBV11KNGTF8N_-fdfmjTZ3WzPQFbLfVrl5n6R8s/edit?usp=sharing




https://i.imgur.com/v4xu4s1.png - OLD DIFFICULTY from December 6th 2017

New from January 26 2018.
https://i.imgur.com/2oBJ4ZD.png