Quote from: Vexual on September 16, 2013, 02:39:16 AM

Quote from: Vycid on September 16, 2013, 02:11:43 AM

https://bitcointalk.org/index.php?topic=99497.msg2788821#msg2788821

Using the sale price of the hardware (which was known), it is possible to determine how many blades were manufactured for sale based on the sale proceeds. Hardware deployed is known: 50 TH (FC did disclose this number, look through his post history). Summing these two and dividing by manufacturing cost gets you a cost per hash figure.

I have no interest in finding the scratch paper where I did this nor in typing it up. Do it yourself.

As mentioned, USBs are a confounding variable and require some guesswork. If their manufacturing cost was significant, then $1.50 is too low an estimate, which I am obviously comfortable with.

So do we take mining income as displayed on Friedcat's July 23 Update as being profit after expenses too?
You have no way of relating expenses listed on that update to the expenses directly related to the hardware sold over that period.
You also dont know what price some of the hardware was sold at, for example, cost to resellers and direct bulk buys.
This 1.50/GH number that your calculations rely on could be out by some order of magnitude.

Obviously mining income is ignored, we already know the size of the farm. We are only interested in the total hashrate produced in order to determine a cost per gigahash.

It could be off, but guess what? All this would make per-GH costs higher, not lower.

I've established a best-case scenario for AM and it STILL appears significantly overvalued. Thus my short position.