I think there's just been a cockup on BTC raised in each line (and not anything dodgy going on) - as it appears from the table that 4 million shares at .0016 raise the first extra 7.5k public BTC then 3 million more public shares (at the same price) raise another 7.5k BTC. When in fact neither of those sizes of public share sales raises anywhere near 7.5k. It looks like the table was changed at some point and one of the columns (BTC raised) didn't get updated when the others did.
The examples in the table are for illustrative purposes and the numbers are correct. The public/private share ratio scales down/up such that a public share represents 10MH/s at the outset of the farm. The "BTC raised" in the table represents the capital expenditure for hashing power only and not the setup/logistics costs involved in deployment of the ~6 metric tons of actual mining equipment, which is covered by the 4400 BTC mentioned above.
See:
http://i.imgur.com/Usl84Ah.png (as used in the BTC-TC listing), it includes the public/private share split.
Sorry - but the math just doesn't add up. Let me explain why in more detail.
At the minimum (current level) you say:
20k BTC raised from public sales is being paid to Hashfast (on top of that there's 4400 BTC being used for infrastructure).
If it reaches the next tier your table says:
27.5K BTC raised from public sales is being paid to Hashfast (presumably the infrastructure costs are at least as much given there's now more hardware to be deployed).
So 7.5K more BTC from public sales is being paid to Hashfast.
But your table shows only 4 million more public shares having been sold. 4 million shares at .0016 is only 6400 BTC. So even if all shares are sold on the market with zero fee you just aren't actually 7.5k more BTC from public sales.
The same problem exists but even worse for the last batch of shares.
I can think of various explanations for this and the NICEST one is that there's an error in the table (either in the numbers or in the labelling of the columns). One such explanation would be that private investors are only putting cash in if the extra batches sell - and that their cash is listed as being as "Raised from Public" which is clearly wrong (and which would explain why your explanation of the numbers appears to include $0 from private investors if there are no future sales). So the answer could simply be that the column titled "Raised from Public" should be titled "Paid to Hashfast" and includes payments from private investors that depend on the extent of future sales.
