AntMiner S9 rampant speculation thread, have to be quick though

[suchmoon]

Bitmain confirmed that the S9 is coming soon™:

http://archive.is/RANno#selection-4047.0-4109.5


Edited 2020-11-28 to fix a broken image

So obviously the right thing to do now is to speculate on the specifications, delivery date, price etc. We have to do it quickly though - before the actual announcement - because it's no fun to speculate on something that is known.

Anyway, my guess is 10 TH/s, 1500W (they tend to creep up with each release), two weeks before halving, and too fucking expensive.

Edit: it's official - 14 TH/s, 1375W, shipping starts June 12, $2100 (~ 4 BTC as of May 31, 2016)

https://bitcointalk.org/index.php?topic=1493601.0

[OgNasty]

Guesses on S9 specs?  9TH/s ~ 1,300 watts ~ $2,200

or more realistically 8TH/s ~ 1,350 watts ~ $1,599

[suchmoon]

Guesses on S9 specs?  9TH/s ~ 1,300 watts ~ $2,200

When do you think it will show up? I think there are two opinions here:

1) before halving - to take advantage of fools who look only at the current profitability
2) after halving - for Bitmain to mine for themselves at the current profitability

[fanatic26]

random guess based on zero useful information


10T @ 2300w for around $1800

[Erumara]

Gonna guess 6.31TH @ 1380W, hopefully in the 5BTC price range!

When do you think it will show up? I think there are two opinions here:

1) before halving - to take advantage of fools who look only at the current profitability
2) after halving - for Bitmain to mine for themselves at the current profitability

I think you're absolutely correct on both opinions. Why would they do one or the other?

[OgNasty]

random guess based on zero useful information

10T @ 2300w for around $1800

I wouldn't buy that now.   

When do you think it will show up? I think there are two opinions here:

1) before halving - to take advantage of fools who look only at the current profitability
2) after halving - for Bitmain to mine for themselves at the current profitability

I am guessing around May 21st.  I believe it will go on sale before the halving because 1) profitability calculators & 2) collecting BTC before it potentially goes "to the moon" after the halving.

[Biodom]

^^^^they might ask >$2000, but there will be very few takers considering that you can have the same hash for $1200.
they might want to hit 10K as they probably wanted to get to 5k in S7, but 9.3-9.5 is more realistic for 1400-1450w miner, could even be 8.6Th and 1293w. 1500 is too close for comfort and reduces the US market.

[Biodom]

random guess based on zero useful information

10T @ 2300w for around $1800

I wouldn't buy that now.   

When do you think it will show up? I think there are two opinions here:

1) before halving - to take advantage of fools who look only at the current profitability
2) after halving - for Bitmain to mine for themselves at the current profitability

I am guessing around May 21st.  I believe it will go on sale before the halving because 1) profitability calculators & 2) collecting BTC before it potentially goes "to the moon" after the halving.

i agree, posted in main thread about this.

[adaseb]

I think it will use the same amount of power as before around 1200-1300Watts but be more efficient so 7TH/s.

Release within 2-4 weeks.

Ognasty was right, try to get as much BTC as possible now before the halving.

[Biodom]

I think it will use the same amount of power as before around 1200-1300Watts but be more efficient so 7TH/s.

Release within 2-4 weeks.

Ognasty was right, try to get as much BTC as possible now before the halving.

perhaps, but only batches 1-7 were in that range (1100-1300); later they crept up the power usage.

[tntdgcr]

I think it will use the same amount of power as before around 1200-1300Watts but be more efficient so 7TH/s.

Release within 2-4 weeks.

Ognasty was right, try to get as much BTC as possible now before the halving.

I like this mindset!

[adaseb]

My guess is that it also won't be more efficient then ETH mining currently, unless Bitmain can make it 8TH/s and use less than 1000Watts. But that is almost 0.10W/GHS territory.

[Prelude]

My guess:

8TH/s, 1200w. Release ~2 weeks before halving to take advantage of suckers. $1200

My guess is that it also won't be more efficient then ETH mining currently, unless Bitmain can make it 8TH/s and use less than 1000Watts. But that is almost 0.10W/GHS territory.

What's your point? We're talking bitcoin mining here, not an altcoin. There's a separate forum for that.

[suchmoon]

Gonna guess 6.31TH @ 1380W, hopefully in the 5BTC price range!

I hope not these specs - not much better than S7 and 5 times more expensive. Although batch 1 might be horrible as usual.

[notlist3d]

I thought it was going to turn out to be trolls on this.  But if they are getting official accounts... that changes my mind.  I have no idea what to expect though.  I think design wise will be like S7 with 3 hashing boards.  But specs.. who knows.

[Finksy]

I would be willing to guess we are looking at 0.09W/GHs ATW efficiency, 14TH/s and around $1900.

[suchmoon]

I would be willing to guess we are looking at 0.09W/GHs ATW efficiency, 14TH/s and around $1900.

I would love that, perhaps would even buy batch 3 or 4. Bitmain tends to improve power consumption by 30-40% with each generation. Going down by 70% (0.30W -> 0.09W) would surprise me.

[64dimensions]

BM is constrained.

Back of the envelope:

9 TH/s

1300W

2.6 BTC

1) 1300W with a useful tolerance means a 125V 15A receptacle can be used.

2) 9 TH/s is a minimum.

3) 2.6 BTC is a maximum



Depending on your opinion of ETH, 2) and 3) can be influenced by ETH mining.

[Finksy]

I would be willing to guess we are looking at 0.09W/GHs ATW efficiency, 14TH/s and around $1900.

I would love that, perhaps would even buy batch 3 or 4. Bitmain tends to improve power consumption by 30-40% with each generation. Going down by 70% (0.30W -> 0.09W) would surprise me.

Efficiency figure is pure speculation based on an exponential formula I worked up in Sept '14, but so far one of them has been relatively close to predict S5 and S7 efficiency. As for pricing, it's a complete shot in the dark.  You never know with BMT...

I can't imagine they would change their form factor for this generation, after the success of the S5+ and S7.

Just thought I'd try to contribute something.  Your model relies on a fixed 1 W/GHs which we know is not realistic, so I thought I'd try to make a model to predict future chip efficiency using past and current developments of both BitmainTech's and Avalon's chips, as they are 2 of the bigger players that have been around for a while.  Please feel free to interject where you see any of my assumptions as unrealistic:



Both Exponential formulas come out somewhat closely, so I feel it could be useful to predict future technology as far as efficiency goes.  We know there has to be a point of diminishing returns.

Interesting to revisit my post from this long back as an estimate to the S5, S7 and future Bitfury chip efficiency.   

When the S5 was released, x=25 for Jan '15,

11.2e^(-0.118)(25)=0.586 W/GHs
7.0376e^(-0.112)(25)=0.428 W/GHs

When the S7 was released, x=32 for August '15:

11.2e^(-0.118)(32)=0.257 W/GHs
7.0376e^(-0.112)(32)=0.195 W/GHs

Let's say we see Bitfury chips by April '16, or x=40

11.2e^(-0.118)(40)=0.0998 W/GHs
7.0376e^(-0.112)(40)=0.0798 W/GHs

Could be on to something here, first formula seems to hold somewhat true at this time.

[DOGE12321]

It will most probably have to still be profitable despite the coming halving, so I'd say 8-9 TH/s with a $1.3k price tag. Maybe super efficient, 0.075w a gh? (Hopefullness)