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.

Technology Stocks : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?

To: pirasa2 who wrote (238295)	8/7/2007 1:44:01 PM
From: dougSF30	Respond to of 275872

By Q4 they will have both Barcelona and Phenom. No. "by" Q4 they will have Barcelona. "by" the end of the year, they will have Phenom launching with a few models, in the words of Henri & Dirk, "so late in Q4 as to not meaningfully impact revenues". From a production stand point, 1 million finished products in Q4 is doable from 200-250 300 mm wafer starts per week during the 13 weeks of Q3, and that's afer assuming horrible yields. You are off by an order of magnitude. AMD will generously get 60 Barcs per 300mm wafer. 60 * 250 * 13 = 195K, not 1M ! If these multiple errors are the basis of a long investment in AMD, you'd best reconsider pronto. I mean, you also believed they burned 500M in cash in Q2, when in fact it was 1.2B.

To: pirasa2 who wrote (238295)	8/7/2007 2:25:52 PM
From: graphicsguru	Read Replies (1) \| Respond to of 275872

I don't doubt that AMD has the manufacturing capability to build a million Barcelonas. They just couldn't sell them. See wbmw's post. I think he's in the right ballpark. That implies a brief peak of around 800K opterons per Q, back down to about 400K/Q right now. Note that the 800K number was achieved when AMD had a significantly better product than Intel for most purposes. With Barcelona launching at 2Ghz, against 3Ghz Clovertown, that's certainly not true this time around. More to the point, the single thread performance of 2 Ghz Barcelona will be significantly worse than 3Ghz Opteron for almost all purposes, so it won't even be a slam-dunk upgrade for existing Opteron users. Not all server apps are about throughput. Latency is important for many purposes. At 3Ghz, Barcelona would be a pretty interesting product. At 2Ghz, it's a niche product. Until the clock speed gets competitive with Intel, I expect Barcelona volume to stay under 100K/Q. AMD will win a few big HPC contracts, but most people will wait until the clock speed gets a whole lot better. And by the time the clock speed gets more reasonable, Barcelona will be competing with Penryn at faster clocks, better IPC and larger cache. As far as Phenom goes, I'm really not at all sure why many desktop users would prefer it to a higher clocked A64 x2. Certainly not for a premium price until its single-thread performance beats the x2. And remember how aggressive Intel's desktop pricing already is. Phenom isn't going to do any volume unless it's priced very low. I think AMD will make more revenue building multiple x2's in the short run.

To: pirasa2 who wrote (238295)	8/7/2007 3:16:07 PM
From: wbmw	Read Replies (1) \| Respond to of 275872

Re: From a production stand point, 1 million finished products in Q4 is doable from 200-250 300 mm wafer starts per week during the 13 weeks of Q3, and that's afer assuming horrible yields. Your math needs correction. Assuming AMD's explicitly stated yields of 0.5/cm2, they can only get 60 working Barcelona die per 300mm wafer. That would therefore require more than 1200 wafer starts each week to ramp to 1M. Besides which, I already showed that AMD's current share of the server market is around 400k units (in Q2). If they doubled their market share (which they won't with 2GHz Barcelona), they still won't get to 1M units. Re: The global server market is 15+ million units No, it's less than half that. I found the total market for servers is about 2-3M servers per year in a post I made for Petz. We multiplied out the number of likely processors, and then divided by 4 quarters. You can probably search for this post better than I can with my limited account privileges. Re: Surely at its height in 2006 AMD achieved better than 25% unit share in a quarter They did, but their most recent quarter of share amounted to 12.8% of the market. Barcelona won't win this back in 2007.