https://bitcointalksearch.org/topic/m.9425492
Also, the Prisma may not have been a pre-order because the hardware was already designed, but the BE200s were not in-hand, and when the first batch came back flawed they were unable to react (let alone communicate) fast enough, so I'd ding them again for handling a hardware launch extremely poorly.
I agree communication has been absolutely terrible. TBH I think it is significantly impacting the company's sales.
The BE200's were in hand, just many of the chips were defective.
Do you have a source on the chips being in hand but defective?
The best source I have was the post on October 24th from friedcat in a different topic I cross-posted it to the Prisma topic here:
https://bitcointalksearch.org/topic/m.9436661
They gave similar details on their Facebook page around the same time. No other details have been released from friedcat that I've seen, and phasebird's only reply was in the Prisma topic on the 25th:
https://bitcointalksearch.org/topic/m.9324615
It sounds like either TSMC had damaged some of the ASICs, or (more likely) there's a defect in the way they sent the VHDL to TSMC and their tapeout went...poorly.
It's enough that they aren't doing another run of BE200s, and friedcat jumps into describing the BE300 and its power efficiency and timeframe as being Q1 2015. So I think the BE200 issue is on ASICMiner's end, not on TSMC's.