@ tarmi, empoweoqwj, ninjarobot
Guys, you are so wrong...
Look, did you remember last time we discussed blade production costs?
(actually that's a recurring topic, like many in this thread; I guess today is just the time for yet another iteration)
So we concluded that the mass production costs for a blade should be like around $20. That's a reasonable price if you compare it to consumer products of similar complexity (like graphics cards). Remember it's China and Friedcat's prices are likely very close to the lowest available anywhere in the world, at least now that the volume became so high. Okay, let's be very pessimistic and assume a ceiling of 30$, that's around 0.25BTC by current prices, and with a reasonable profit margin it would still be more than acceptable for them to sell the blades for less than 0.5BTC. Now please reconsider if people would buy these blades in November, if they sell for 0.5BTC! I bet they would even in December, even if most the competitors actually deliver in time with their currently announced prices.
Also note that most competitors are less likely to have production costs as low as Friedcat (you know, Chinese are very social people; the "connections" are very valuable for them so a foreigner cannot get as good deal as a local).
Also note that 55nm chip of the same die area is much more expensive than 130 nm chip. Of course, it's still cheaper per Gigahash, but not as much as per more powerful and efficient. Let's say, if the 55nm chip is 8x as powerful as 130nm, then it's likely about twice as expensive per chip, so only 4x as cheap per GH.
So don't underestimate the good old 130nm, it will last at least until the end of the year.
(EDITED: "at least")
500 TH = 50 k blades = 5 k blade kits
I am not sure that you are aware of logistic and scalability problems of deploying 500 TH of asic blades. Not to mention power consummation and cooling.
They dont have gen2 chips, they are late, and are hoping to recoup that with old technology.
