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: Paul Engel who wrote (57181)	5/4/1999 1:50:00 PM
From: Scumbria	Read Replies (5) \| Respond to of 1572881

Paul, With June only four weeks away, your attempts to demean AMD seem increasingly trivial and dull. Scumbria

To: Paul Engel who wrote (57181)	5/4/1999 4:18:00 PM
From: A. A. LaFountain III	Read Replies (4) \| Respond to of 1572881

Re: "sitting on 900,000 to 1,000,000 K6-2 parts" I'm intrigued by your statement and would be interested in the source, since there was zero indication of inventory accumulation in the 1Q financials: $(MM) Days 1Q97 $149.2 39 162.6 40 163.4 35 168.5 36 1Q98 157.9 34 171.6 40 172.2 37 175.1 33 1Q99 182.7 37 If the inventories have ballooned since the start of the quarter, that would imply that 200,000 units have accumulated each week over the five weeks of the quarter that have elapsed. This is a very unappealing scenario. But I'll confess to being confused. If I follow the gist of your threads, it has been that AMD has been unable to make sufficient quantities of high speed parts to be a profitable vendor. Management has indicated that the manufacturing problem was resolved (with a vengeance) during the first quarter, so the production of high speed parts has moved up significantly over the past ten weeks. So if there has been an accumulation of inventory during a period in which AMD has had high levels of production of high-speed parts, either (1) the PC market has cratered; or (2) Intel has continued to pick off share. If the answer is (1), then all the finger pointing at Compaq over the last two weeks for getting people to believe that it was a PC industry problem in 1Q when in fact it was just a CPQ-specific problem could very well be misplaced, and there is an industry problem. If the answer is (2), then the claims of some people on this board that AMD's problems are all of its own doing would now appear to be misfounded. Personally, I lean towards (2), with the acknowledgment that Intel has indeed become a much more rigorous competitor at the low end. Where you and I might disagree is where this leads. I believe that Intel's strengthened presence at the low end has already begun to completely cloud the 4-segment marketing approach (anyone who buys a PIII system at this point must feel that Intel needs the cash more than they do). More importantly, down the road, this all becomes virtually irrelevant, as the K-7 eventually goes after the workstation and server market. Now, I know we are discussing a paper (foil) tiger here. But I keep coming back to what I believe is the key element of the story - the Xeon pricing structure. If I understand the market reports, the Xeon sells at $950, $1,850 and $3,950 for 512K, 1M and 2M cache versions. Unless I'm missing something, that means Intel is selling SRAM at about $2,000 per MB and is throwing in the processor for free. Of course, that's a ludicrous notion (even though the arithmetic supports it). So it means that Intel has in fact documented enough of a performance enhancement with the higher levels of cache that the customer really has no alternative but to cough up the extra bucks for the higher-priced parts. But there's a big disparity between what Intel charges for these parts and what they cost. If, as I suspect, Intel is enjoying 95% gross margins on the incremental SRAM packaged in these parts, then we're confronted by the possibility that a significant portion (maybe up to 20%) of Intel's $9 billion in operating profit from the Intel Architecture segment in 1998 was generated by SRAM sales. My guess is that very few investors in this microprocessor vendor realize how much of the profit stream is generated by lowly SRAM (and more power to Intel for being able to turn SRAM into gold). But the K-7 entry into this segment puts this profit at risk. If this is the case (or if the case is anywhere near this analysis), then it would be a mistake to focus on AMD's well-publicized shortcomings of the first quarter. Attention to AMD's problems over the past several years as it has moved into the MPU business with its own designs does lead to a sharp differentiation with Intel's record. Yet to assume that the past is anything more than prologue could lead to a substantial investment error. - Tad LaFountain

To: Paul Engel who wrote (57181)	5/4/1999 4:33:00 PM
From: Gopher Broke	Respond to of 1572881

AMD is reported to be sitting on 900,000 to 1,000,000 K6-2 parts slow & FAST - that they are having trouble to unload Please substantiate.

To: Paul Engel who wrote (57181)	5/4/1999 8:12:00 PM
From: Cirruslvr	Read Replies (2) \| Respond to of 1572881

Paul - RE: "And you should worry - AMD is reported to be sitting on 900,000 to 1,000,000 K6-2 parts slow & FAST - that they are having trouble to unload." Didn't someone say there was a possibility AMD wrote off in Q1 earnings the low MHz K6-2s that weren't sold yet and that were made during the mask problem? From the Q1 press release: "The results for the first quarter include a charge of $15,016,000 for restructuring and other special charges." amd.com That could possibly take care of the slow chips you speak of, but not the FAST ones. AMD has always had strong demand for the K6-2. Why would that demand start to wane all of a sudden? From my biased perspective, I don't see that happening. Also, in this message by you Message 9317507, you said this - "My source - which is "third hand" - comes from an AMD source in the Pacific Rim area. Apparently, orders for these devices are slowing to a trickle. You may want to see if you can spot some low-ball K6-2 pricing in that region as AMD attempts to unload these parts at very low prices - before they become "obsolete"." Assuming you are right about the region and that AMD has 9E5 - 1E6 K6-2s they are trying to unload leads to an assumption that AMD has a LARGE portion of its business in the Pacific Rim area. I can't find all my notes from the Q1 CC, but I believe someone asked what AMD's breakdown of overseas sales was (I think it an overall number, but since chips make a great portion of their revenues we should be able to work with it). If anyone has that number (Tad may possibly have it), we can see if Paul's number jives in any way with what AMD says. I personally think 9E5 - 1E6 lame K6-2s seems too much for that region. That range would only make sense if demand there was practically ZERO.