Maybe your calculator is doing something foolish, like the genisis version does. Difficulty in the trillions early next year? Not likely...
When the Bitfury rigs no longer make money, nothing on the market or anticipated to be on the market this year will, either. They are as efficient as some of the up and coming 28 NM chips...
That the calcator I'm using. So what you are saying is that a hashrate that high isn't realistic. What are you predictions then for a return on a 400 gh/s unit.
I think past ~december, the monthly increase (currently around 76%), will taper towards 50%, and by ~february/march should be a lot more reasonable, maybe 35%.
right now, GPUs are pretty much all stopped mining bitcoins
soon, FPGA will follow (~45w/Ghash) at $0.10/kwh, these are only making about $0.35/Ghash/day right now, and in another 2 months will be costing more electricity than making BTC
next, avalon will stop (~10w/Ghash
then, asicminer will stop (~8w/Ghash)
then BFL (~6w/Ghash)
bitfury is operating at <1w/Ghash, which puts it in line with 2nd gen ASIC designs, and enables them to mine at a profit for MONTHS (years?) longer than the above competitiors, which currently account for most of the network