Re: AntMiner S9 rampant speculation thread, have to be quick though
If they take orders on May 22 the soonest it get to usa is June 10th
How do you figure that? I mean if it's immediate delivery then it would get there within a week, but if it's a pre-order then it can be any lead time, about a month or more based on past performance (e.g. S3 and S7).
well they do not want to toss all their psu's away.
Agreed. But I would add that they probably don't want to release a design that would force any significant changes to their existing data center infrastructure.
Therefore the S9 will likely be a drop-in replacement for the S7. In addition to not overtaxing their existing PSUs and not requiring new cabling, it will likely be the same form factor. If true, that would be good news for those of us who already have a PSU infrastructure geared towards S7s.
I think the same was said about S5 and S7 yet both miners ended up significantly different from their predecessors. The only constraint now is that they are approaching the limit of 15A household circuits in North America (~1450W or so) if they care about that.
I have no doubt that Bitmain will try to offer sales in May so that they can cut BW off at the pass. I think that BW said that they won't offer their 14 nm miners till June. But I wouldn't be surprised if that slips past the halving, well after they've sucked out all the sweet pre-halving difficulty that they can.
Regardless, BW's publicly announced plans offer Bitmain another opportunity to pounce and conquer, if they can execute well. Not clear if they will compete on efficiency or cost. Although I suspect that Bitmain doesn't have a smaller nanometer process ready to go. That would suggest that probably trying for another efficiency gain with the existing process, as they did with the S7.
This would further suggest that they can dramatically undercut BW's prices and still be profitable, so it might be wise to not buy the first batch of S9s (as if it's ever a good idea to be the first penguin off the ice flow in this market; my B1 S7s definitely got eaten by the killer whale called "difficulty").
The main open questions are how efficient will the S9 actually be, and what will it cost? Although whatever it costs, it will be too much
Regardless, BW's publicly announced plans offer Bitmain another opportunity to pounce and conquer, if they can execute well. Not clear if they will compete on efficiency or cost. Although I suspect that Bitmain doesn't have a smaller nanometer process ready to go. That would suggest that probably trying for another efficiency gain with the existing process, as they did with the S7.
This would further suggest that they can dramatically undercut BW's prices and still be profitable, so it might be wise to not buy the first batch of S9s (as if it's ever a good idea to be the first penguin off the ice flow in this market; my B1 S7s definitely got eaten by the killer whale called "difficulty").
The main open questions are how efficient will the S9 actually be, and what will it cost? Although whatever it costs, it will be too much
I wonder if the S9 is really going to be a 14-16nm miner. Can Bitmain attempt to squeeze some efficiency out of their current chips? According to the BM1385 datasheet power consumption can go down by about 30% with undervolting.