Guys seriously you will be lucky to get a 5970 before January. Every 5870/5970 is going to the high end pre-built systems.
TSMC are lucky to hit 10,000 wafers a month when it should be well over 20,000. Add to this the low yields and pressure from the likes of Dell...
They can only do 9k 40nm waifers a month, and they are down to 6k till after xmas, thats for both Nvidia and AMD for all their ranges(well Nvidia low, low, low end, AMD mid and high end).
The reason you probably didn't get many 5970's doing much end of last week, thanksgiving and a lot of uk retailers buying supply off American distrubutors(well quite often, hugely bigger market, will often have more supply than a smaller distro just for the uk). Imagine shipping time is, normally 5 days USA to being in stock in the UK store. They buy, they put up the ETA and they could easily forget that customs/shipping could easily take a few days delay over thanksgiving. I was watching the US stock market on Thursday and couldn't work out why it wasn't opening, Thanksgiving is just something no one remembers as we don't have it here, I'd bet thats the main issue.
AMD have stated mid december for supplies to ease up. I would assume it because, well waifers take several weeks to go from start to finished, one of the biggest problems is TSMC's problem wasn't caught during production, they only found out basically when you would have had a huge batch of waifers ready to ship and found they didn't work
Which means they have to wait for the next large batch to come out of production before they can start shipping in larger numbers again. Its NOT like a slice of silicon goes in the plant in the morning and comes out in the evening as cores that ship the next day to AMD.
THe dates company give for stock supplies in this case, AMD saying when supply will ease up, will be accurate because they know when a big batch of waifers is due.
Lets just all pray, cross out fingers and do some kind of high yield/TSMC not screwing up dance to the gods and seriously hope TSMC haven't screwed up another huge batch of cores. Supposedly only one production line had a uncalibrated part so the other lines "should" be fine.