That seems to indicate they aren't going to be there with their next chip, unless something changes.
That is an excellent timetable. With the faster specs will heat be more of a problem?
Guy: No, itll have the same package and thermal as the 2nd gen but at a much higher hash-rate an much reduced cost

Its a design to compete and win in the 1,000 PH/s era (2015) We want to be able to offer 0.2$/GHs solution to our customers with 0.1W/GHs. We are not there yet, but Im confident in our team ability to achieve it on 28nm.
If in October they haven't got the simulations to 0.1J/GH, I wouldn't be hopeful on them getting silicon by the end of the year at that level. If they can though, make props to them. That's 3.4x more efficient than their current chip on the same process node.