...
Now there's several good reasons why the difficulty will -not- get to 7 billion with the price of bitcoin at $125. There's ROI to think about, the cost of hardware must be amortized in, which means the difficulty is actually very likely to level off at more around 2 billion or less.
...
You're ignoring that:
1. Miners pre-order their gear months in advance, so the gear ordered today is *locked in* -- it will mine even if it makes pennies over power cost.
I'm not ignoring that. I assume people have pulled back on their mining equipment orders in the last months as diff skyrocketed. If I were looking to buy today, I wouldn't.
2. People are bad at math -- they will continue to come up with "but bitcoin will be worth moar" arguments & mine at a loss.
I don't think that's as widespread as you think. People with real money tend to do their due diligence, or else are quickly stripped of capital.
3. Even when people are given a chance at a refund (the case with KNC?), they don't use it -- even when all evidence tells them to do so. Some yet unnamed gambler's fallacy, i guess.
4. People bought & are still buying USB miners -- case in point.
When the cost is a few $10 bills, no one cares that much if they make money. I look at that as the chewing gum of ASICs. You never need a reason to chew gum. But a Jupiter is a 6 course meal.