My mistake, my calculation was based on a 2 week block time and 30% *hash rate* increase per 2 week period, not 30% diff increase.  I still don't believe we'll maintain the growth rate that started in august.  I stand by my prediction of 1.5-2 billion by march.  We would have to see a number of mining hardware manufactuers over deliver to get much past that.

Based on delivery estimates vs actual delivery for many of these companies, I think the rollout of the next generation of chips will be slower than many are hoping for.  After this jump to 28nm chips, we'll continue to see the growth rate slow as that's almost caught up to mainstream chip technology.

As I've said in other posts, ASIC chips have effectively gone through 7 years of chip processing advancements in less than a year.  That partially explains the massive increases in hash rate.  There was an ASIC gold rush, but its just about over so we won't be seeing these 30-40% jumps as a regular occurrence for much longer.

Eventually the growth slams up against equipment demand (dropping as projected ROI drops) and limits of current chip technology for efficiency in hashes/s per watt and $(BTC)/hashes/s.