Quote from: kleeck on October 31, 2013, 07:03:54 PM

Quote from: Ozymandias on October 31, 2013, 06:18:19 PM

Quote from: kslaughter on September 19, 2013, 10:12:00 PM

What are the current plans for chip development?
ActiveMining is developing a 28nm Bitcoin mining chip using eASIC.

Chip samples delivered in 9 weeks;
Low-volume chip production starting in 12 weeks, using an e-beam process;
Normal volume chip production starting in 16-18 weeks.

So if we set our start point as a week before September 4th (when the eASIC press release occured), we would now be in exactly week 9. If this is in fact the case, then low-volume production should happen in the third week of november and normal volume around the holiday season of december. If we're good, Santa (or whichever obese, brightly colored man breaks into YOUR home in the dead of night in the winter) might leave a little extra in our stockings (or similar footwear).

Any asic specialists want to inform the ignorant, myself included, about what kind of volumes are considered "low"? A few hundred? A thousand?

For reference, with a current network hashrate of ~4ph/s, it would take roughly 2000-2500 chips per percentage of the network. Obviously difficultly is not stagnant, and is in fact doubling a little faster than every month. With a 16ph/s network around the new year (2 doublings), each percentage of the network would require 8000-10,000 chips, or (iirc there are 2,500 chips to a standard wafer of 28nm chips), we would need a mere 40 wafers to corner 10% of the market.

After some more basic research, I'm finding that a conservative estimate for the cost of a wafer is ~$4,000. Thus 40 of them would cost $160,000 which is just about HALF of what the avalon refund was worth (~1500btc*200usd/btc = $300,000). What I'm seeing, if all of my assumptions are correct/close (which is an if I certainly wouldn't bet the farm on) is that we can order ~100 wafers using our avalon refund and pre-order funds.
100 wafers = 250,000 chips * (0.95 to account for defective chips) = 237,500 chips * 16-20gh/s = 3.8-4.75ph/s to be distributed between our own mining farm and hardware sales.

If any of the above is wrong I'd welcome corrections.

Here's what I got:

16gh per chip
256 gh/s per board
16 chips per board

40,000 (1%TNH@4PH) / 256 = 156.25 (boards needed at 4PHTNH)

156.25 * 16 = 2500 chips. So, yep. Checks out. The question now is, "How many can we afford?"

$1.5 million worth