So some calculations on this IPO.
The OP's post #4 claims 6000-10000 1 TH miners produced within 1 year. Taking the lower estimate of 6000 TH produced in one year, let's do some math:
Given current difficulty rates are approximately doubling every month, and assuming hardware is produced on a continual basis (say, 500 TH / month to get to the 6000 TH in one year), then the BTC mined by the hardware will be:
2600 BTC for that sold in 1st month
1300 BTC for that sold in 2nd month
650 BTC for that sold in 3rd month, etc
For a total of 5200 mineable BTC.
To determine the amount of money payable to shareholders as dividends, one must subtract the costs of the chips and any other expenses. Let's assume that friedcat charges them $0.50 per GH as he mentioned in his chip thread here:
https://bitcointalksearch.org/topic/m.4816701 For 6000 TH, that's 3 M USD, or about 5000 BTC at current prices (~$600). Even assuming there are no other costs (obviously not true), that leaves only 200 BTC in profit, split among 75000 shares, or about 0.0027 BTC/share.
Even if friedcat offered the chips to RM for half the price he promised (and thus hurting his own investors), $0.25 per GH, that means 2500 BTC in chip costs. Again, assuming no other costs, that leaves 2700 BTC in profit divided among 75000 shares, which is 0.036 BTC/share.
In reality, costs other than chips are likely to be large, likely more than the chips themselves given the realities of every other industry I've ever heard about. The PCBs, miner boxes, employees, space, overhead, etc, all cost plenty of money, reducing any above mentioned profit further, or in fact even resulting in a loss.
This seems like not a very good deal for investors, unless one assumes they are buying into just the start of a business which will then grow far beyond its original plans (but this is not claimed/stated in the OP). Unless RM builds a lot more miners than projected, or plans to sell them to customers for far more than they'll ever mine, or difficulty growth slows substantially, 0.1 BTC per share seems like a lot.
Thoughts? Any mistakes in the above? I did this pretty quick so could be wrong on the numbers somewhere.