There is a little new information in Ravi's quote:
Another key thing to note, added Iyengar, is that our GoldStrike1 ASIC itself is very high performance and meets and exceeds the published spec. It is the board level power circuit inefficiencies which reduced the overall performance marginally. However, as stated above we are getting closer and closer to the spec with improvements that we are doing on the board."
So perhaps they were accurate when they reported their 132 GH/s/die. If true, I'm guessing the tinkerers out there will be able to get plenty of further performance out of the under-speced units, much like October KNC units.
of course it's possible, but highly speculative, that they have decided to do a quick board redesign before they ship the first December units - I think the turnaround on a new board could be just a few days if they really act quickly. This would explain the delay in shipping the first units. This is just complete guesswork though.