My speculated fair value: 6.3 BTC per share.
wow, that's a whole lot of optimism!
They won't be getting the same value for the blades as they were before, mainly because of competitors, and lowering ROI.
Also, mining revenue (as we have just discussed) will probably be decreasing. It doesn't matter that they have 200TH, because 200 TH in 2 months is not the same as 200TH today. Instead, think of it as a % of the total hash rate. If they can keep 20%, then we are looking at consistent .015/share per week.
Optimism has worked out pretty well so far.
The estimate takes into account decreasing mining revenue, but may be on the optimistic side.
Let's do another one. First assume that selling hardware and mining generate the same revenue, then assume that AM hardware keeps mining 25% of the network hash rate - because of the first assumption it doesn't matter whether it is friedcat and friends that are hashing it or someone who bought AM hardware, revenue for AM is assumed to be the same in both cases. Say there's 5 million BTC to be mined over the next four years, then it works out to approx. 3.1 BTC per share.
Looking at the competitive landscape, can AM churn out 25% of hashing hardware? Well, they have already manufactured hardware for 46% of the current network hash rate. They have an established production line, great cash flow, no debt, undelivered orders already paid, and a production cost that (I'm pretty sure) outcompetes all the other companies. I am confident they can manufacture at least 25% of all hashing hardware for an extended period of time, possibly significantly higher than 25% in the short term.
So let's say - conservatively 3.1 BTC per share.