Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?

To: Jim McMannis who wrote (46779)	1/20/1999 7:11:00 PM
From: kash johal	Respond to of 1572512

Jim, >Re: Care to make an estimate as to how many wafers it would take to >get the first 50k good K7 die? Is 20% yield out of the question? So >maybe 2,000 wafers at 25 good die/wafer or about a 20% yield. Would >they release with a yield that low? >At $2000/wafer /25 = $80 silicon cost plus say $20 for packaging >that's $100/chip initially. Is this out of line? my 2 cents: Yields may be in the 10-12% range due to huge die sizes. I suspect fab 25 wafer costs are more like $1500 as they are almost at full capacity. So a $100 die cost and $30-50 cache/pkg/test cost is not unreasonable. So a $150 cost is not unreasonable. If performance is good 600Mhz+ in 2H99 they should be able to sell all they can make. In 0.18 their costs should drop below $100 easily. Regards, Kash

To: Jim McMannis who wrote (46779)	1/20/1999 8:16:00 PM
From: Petz	Respond to of 1572512

Jim, with K7 die size of 184, there should be 150 potential die. I don't think AMD would start production until they were assured of at least 20% yield, so I figure 30 good die per wafer initially, so your $100/chip looks like a pretty conservative (i.e., high) estimate of cost. I think your 20% number is about right, because 30 K7's per wafer would probably be sold for a little less than $10,000. Since Jerry said that AMD got MORE than $10,000 per wafer producing the K6-2, I doubt they'd want to do any significant production if yields were still below 20%. If they ramp up to 500 wafer starts/week by the beginning of March, they should have enough K7-s for a launch in early June. (There would be 30,000 completed K7's and 30,000 in the process of packaging.) 15,000 K7's per week starting late May @~$300 would add $5M revenues/week towards the end of Q2, rising with yield. (I assumed that the wafer production takes 10 weeks and packaging 2 weeks.) If they got 75K K7's sold in Q2 at $310 each, the 23M revenue would more than cover the cost ($2,500/wafer * 500 wafers/week *13 weeks = $16.2M) for the entire quarter. Of course, for part of Q1 there's cost but no revenue. Petz