I have a question.
You have several
How come it is economical to buy somebody elses chips to manufacture miners when not so long ago it was decided that it is uneconomical to continue manufacturing own 55nm chips and that project should be cancelled?
It likely isn't. The nature of the chips is purely speculation. It is just as likely that a magic unicorn farted them into existence as it is we bought them as it is we made them. Until we know what those chips are, how they were obtained, and for how much they were obtained we can only speculate.
At the point when the cancellation was made the NRE was already paid and production run started if we are to believe what Ken says. So those chips that were cancelled were going to be practically free since their costs had been already paid for and since the production run was already started that money was truly spent by the factory so no refunds possible. At the time of when the decision was made the money spent already is sunk costs and should be ignored in calculations and any decision whether to continue or not should be made solely based on costs happening in the future from the time of cancellation and in the case of ASICs already in fabrication those costs are very miniscule.
First off it wasn't a production run but a shuttle run (i.e. very few chips just to see if they work vs. tens-hundreds of thousands of chips). We do not have details on the 55nm NRE so again, we can only speculate as to the costs but the decision to cancel the 55nm chips was made during a particularly brutal part of the bear market (a decision that I do not at all agree with, but can understand the reasoning behind) where the price per chip+board+misc. components would [theoretically since we again don't have numbers] outweigh the revenue.
So almost free own chips determined to be uneconomical and project cancelled. Instead Ken has now bought somebody elses chips at a price that most probably represents all their NRE and production costs in full plus profit to whoever is selling them.
Addressed above.
Can somebody explain the logic and economies comparing these two cases that explains why the cancellation made sense and also why this whats happening now makes sense?
Ken is likely the only one that could since he has data that none of us shareholders have.
Secondary question if there can be found some logic and economies where the first question makes sense, how come this explanation was not discovered already when the decision to START the 55nm project was made? That time is also not in very distant past and anything that is known now should have been able to predict at that time?
The decision to start the 55nm project was made when btc/usd was at or near the point of profitability. When btc/used fell, so did the projected revenue for the 55nm chips and thus the project was cancelled (though this is again mere speculation).
Continuing with the own chips would also serve as learning point towards the future 28nm chip which is much more difficult and risky to design considering for example the much higher NRE that means that the chip should be good on the first try as there likely is not enough money or time for a respin.
ActM has retained the engineers that designed the 55nm to do the 28nm. Your point about the learning point would have to be weighed against the losses of going through with the 55nm chip. Perhaps (not my opinion but likely the reasoning used) the losses would have been too great to be worth the learning experience, perhaps it was some other reason based on information to which I am not privy.