I disagree, we will never see 10M difficulty, as long as BTC price stays below $20. You have to remember difficulty is a lagging indicator of BTC price, we already see that difficulty has leveled off, at the current rate, the next difficulty will just be a 10% jump, and then it's possible that we will be seeing a difficulty decrease (IF price stay below $20).
The problem I see in all of this analysis of bitcoin mining is that you cannot predict the price of BTC that you need to sell to pay for your mining costs. I'm mining with 6950s (stock because I ended up with cards I can't unlock the shaders on - oh well
) I'm getting around 355MHash/s each (900MHz which is as high as I can go locked)
Right now at ~$15/BTC and paying .1/kwh in electricity, mining ceases to be profitable at a difficulty of 10000000 for a single stock 6950. That's over 7 times the current difficulty! Are we headed there? Yes we are. Will BTC stay at $15, crash to $1 or go up to $100? Nobody knows. But right now? I'm earning $$$ to pay off my rigs, and at the very least should be able to recoup enough to break even if I sell my stuff used. But if I pay the rigs off, then it becomes a pure equation of power costs.
So will I pay off my rigs? Only time will tell and it all depends on the market and $/BTC.
But lets say difficulty skyrockets in a month's time and profits vaporize. We actually see a drop in network processing. Then BTC jumps to, say $50/BTC. Miners will pile back in.
Bottom line is NOBODY can predict difficulty, GPU advances, or $/BTC. Just like the day traders, mining has risk. As it should be.