So how much is 300TH/s going to cost them to buy and run? I punched the figures into http://mining.thegenesisblock.com/ and it reckons the 300TH/s might make ROI in a couple of months, but it's a gamble because you don't know who else is coming online, the estimates are only at the known current growth rate. If it were me I would simply put the $millions into buying bitcoins, I don't see economies of scale.

I think it's quite funny that KNCminer talk about the ethics of ROI and taking a break until March for the net hash etc. Yet they are happy to support mega farms that totally fly in the face of the distributed hash concept of Bitcoin.

When the Hashfast, Cointerra, Avalon gen2, ASICminer gen2 all dump onto the market, 300TH/s will be a small player. BFL probably  still have 300TH/s of back orders to fill as well on their current 65nm products. The ASIC vendors are releasing too many TH/s per day, that's going to kill their own market, as the net hash is growing way faster than moores law, meaning they will rapidly run out of efficiency gains from technological advances, and the farms will collapse under their running costs. ASIC hardware is a classic pump and dump, it's didn't need to be that way, the hardware vendors have made it so, they could have sustained a sensible rollout for years.