I think even under the worst case scenario (baring bitcoin crash) the machine will pay for itself within a month. Given that even a difficulty well over 1,000,000,000 would still be profitable...
This isn't even close to true. There is a massive amount of hash power coming this summer, before even the first units of this would ship. It will be at least 650 TH/sec, possibly in excess of 1 PH/sec (the variance is based on how many chips BFL can produce of their 68k x ~2.8 Gh/sec chip among others such as additional avalon orders)
And this is only from the 100% public and documented sources (Avalon, ASICMiner, BFL, etc) This doesn't include ANYTHING from Kncminer, unknown sources, people without confirmed ASICs but still in design, etc.
The only people who are going to make it big with ASIC are the ones who got it prior to this summer. Buying ASIC now is like buying GPUs 6 months ago. You'll break even, but it will take many months to do so.
The advantage of asic
now is long-term mining, because electricity prices will force out all other types of miners, it would require an absurd amount of difficulty to make even the 130nm asic's non-profitable @ electricity costs
