Do the maths, people.
I think you're way off base.
AM probably won't even do much mining, as FC said "It is a safe bet that we will be moving more to pure chip distribution".
Like I said,
In reality, AM will sell a good portion of their chips.
So why did you pose those elaborate (albeit wrong) mining scenarios then challenge us to fault your maths? (rhetorical Q)
The problem here though, is that people will be expecting profits like those from gen 1 chips. That's clearly not going to be the case though. As I pointed out months ago when shares were trading at 4+ BTC, AM had a monopoly at the time which is why they could charge outrageous prices for their hardware. As soon as competition came out, the profits from hardware sales plummeted. Everyone thought I was mad back then for saying that but that's exactly what happened.
This time around, AM don't have a monopoly and there's a lot of competition. We don't really know anything about AM's new chips but they won't have as high a hash rate as the 28nm chips. They may have lower power consumption but that doesn't really matter that much for individuals buying a mining rig or two.
So, AM will be releasing new chips into a highly competitive market and they're unlikely to take the performance (Gh/s) crown. Why should people buy AM hardware if they can get better performance elsewhere? The answer has to be because AM is cheaper per hash.
Lets say that AM got 20 Ph/s and sold the lot in one go and distributed the profits as divs. What would those divs have to be like to make you think a share price of 0.4 BTC was good value?
What? So you're saying AM goes out of business? 20Ph and that's that?
in edit: personally I wouldn't mind if FC said no divs, re-investing in development. I hope he has aggresive and ambitious projects in the pipeline.