Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
DSL
http://finance.messages.yahoo.com/bbs?.mm=FN&action=m&board=4687810&tid=amd&sid=4687...
Together SSDOI and HOT sound an awful lot like DSL. When IBM announced its SSDOI and HOT, in September 2003, it claimed a 40-65 per cent improvement to transistors performance over chips fabbed using a vanilla CMOS process. At this stage, it's not clear whether the 24 per cent gain IBM and AMD are claiming today is in addition to whatever boost SSDOI and HOT ultimately yielded or includes that benefit.
Me...
What is that 24% improvement?
Makes sense, thanks.
Like so many other things there seem to be conficting messages wherever you look. A couple of things I would like to come out of the q4 CC are the number, or at least the percentage of A64s/Opterons/AMD64s being made and better information about how the conversion to 90 nm is going. The product mix and conversion are the biggest items I have on my wish list of answers.
Yes, that's why I added the question mark. It really doesn't make much sense for AMD to be converting processors that will be discontinued to 90nm. AMD usually adds the caveat "As market requires". Still AMD's latest roadmap shows the same thing with 130nm 1-8 Opterons and Athlon Mp's possibly extended.
http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_608,00.html
Given the cost advantages and performance advantages of the E0 steppings it's hard to see a need for these products after q205. Q205 looks to be the end of 130nm for all paractical purposes.
Looks like 130nm production ends by end q205?
http://pc.watch.impress.co.jp/docs/2004/1221/kaigai02l.gif
Personally I liked having a high short interest, but it looks like the old short attraction is fading. These short guys took quite a beating in the run up, and probably are less than enthused about the idea of having it happen again. Still AMD remains very volatile as evidenced by yesterday’s movement. Looking back on yesterday it seemed like a herd of cattle stampeded by lightning. Anyway we got rid of a lot of nervous nellies yesterday which should help as we base for the next move up. To me it seems like risks and opportunities are fairly balanced where we are now.
Short interest dropped a lot.
http://finance.yahoo.com/q/ks?s=AMD
It looks like I was getting revenue mixed up with units.
This is from an august Smith Barney talk.
http://www.investorshub.com/boards/read_msg.asp?message_id=3795722
AMD remains on track for reaching 50% of total MPU revenues on AMD64 by year-end, and crossing over in terms of units in Q205.
AMD did remark that it would not be surprised to see its total AMD 64 (including Opteron) shipments reach 20 million units in
2005, which is slightly ahead of our forecast of 19.3 million.
Me...
If I recall correctly AMD will have a capacity of about 50m units when it gets to all 90 nm which isn't going to happen until sometime next year. So we're talking at least 7 to 8m AMD64s in q405, given that ramping will occur over the year, out of a max capacity of 12.5m units. That still leaves AMD without enough capacity as far as I can see.
That's true, it's usually the off brand hardware that has the problems, but I still expect some problems to arise.
I would think gamers would be chomping at the bit to get their hands on a 64 bit O/S from MSFT. Certainly I can't see any of them buying a non 64 bit machine, and given the move to 64 bit games why would any of them buy anything but a 64 bit O/S?
Also, while I agree that Longhorn will probably be the first O/S that makes full use of the 64 bit hardware, given MSFT's penchant for pushing back intros it could well be 08 before we see the product in the market. In the meantime I can't see many consumers buying a 32-bit machine if the difference in cost with a 64-bit machine is negligible. How long do you think it will be before 32 bit has that passe stigma attached to it?
Granted 32 bit will be around for a long time in markets where price is everything, but I don't want AMD doing anything in those markets other than what's necessary to keep INTC honest, at least until fab36 comes online.
I’m really interested in how AMD64 and in particular how A64 is doing this q. That should give us a pretty good indication as to how well sales will go when a 64 bit O/S is available. I still think, starting in h205, there will be more than enough demand for 64bit to soak up all AMD's capacity until fab36 comes online. Especially if AMD64 product is still viewed as being superior to INTC's offerings.
Right now say AMD is selling 8m chips a q, probably closer to 8.5m . If 60% of those are AMD64 that’s 4.8m chips being sold without a 64 bit O/S from MSFT. If AMD can sell another 4m with a 64 bit O/S, AMD is maxed out at 90nm. This doesn’t seem unreasonable to me given the way AMD64 sales have been growing q/q.
Ok, I'll trust you, pending Reseller Mikes comments. If the distributor sells on consignment to the white-box guy how would AMD know. Of course the distributor wouldn't be able to reorder the consigned units until he told AMD they were sold, but that's another problem.
Thanks, how much consignment stuff goes on. Am I right that the amount varies with demand. I've worked in businesses where terms got very flexible if the situation warranted.
Yeah, drivers are usually my biggest problem. I don't expect to see X86-64 drivers for my antique HP720c printer, but I've got a firewall/router and will keep one of the older machines to hang it off of. Still there will be other problems I'm sure of it.
Since we have little visibility into AMD's chart of accounts I really don't know how AMD handles the accounting, but it probably doesn't matter much. The important thing to understand is that OEM sales are immediately recognized while sales into the channel are only recognized when they get passed on through the distributor.
I read something by Otellini, I think, about how fast inventory looses value recently, but I can't remember the amount or time period. Suffice it to say it doesn't get any better with age. It will be interesting to see just how long it will take before k7 chips finally disappear, given the Lazarus like Duron .
No what is semi-finished goods, not a financial statement grouping. Do you mean WIP? Chips do not hold their value which is the primary reason for not having more inventory than needed at any stage. The White box guy may have the right to return to the distributor, eg., the distributor may sell on consignment, so there is no sell through until the White Box guy sells to the ultimate user. The amount of this sort of sale going on usually is determined by demand.
Well, there has to be a certain amount of inventory in the channels. Some sitting in a warehouse, some on a slow boat, some even sitting in the white-box guy’s back room. I don't think AMD keeps much of its finished goods at the packaging plant, or anywhere else since the stuff is not like wine (doesn't age well). As soon as it's finished it's probably shipped to someone and then sits in the channel for how long?
Anyway, since this is something that will only be proven with time we'll have to wait, but it is something to consider.
As far as the MSFT O/S goes, I think you're wrong and there is a lot of pent-up demand just waiting for MSFT to get its' act together. Not only from business where I expect a very warm reception, but also from lots of little people who don't want to buy a machine with XP loaded and then have to reload the 64bit O/S. I'm one of those people.
Well, there has to be a certain amount of inventory in the channels. Some sitting in a warehouse, some on a slow boat, some even sitting in the white-box guy’s back room. I don't think AMD keeps much of its finished goods at the packaging plant, or anywhere else since the stuff is not like wine (doesn't age well). As soon as it's finished it's probably shipped to someone and then sits in the channel for how long?
Anyway, since this is something that will only be proven with time we'll have to wait, but it is something to consider.
As far as the MSFT O/S goes, I think you're wrong and there is a lot of pent-up demand just waiting for MSFT to get its' act together. Not only from business where I expect a very warm reception, but also from lots of little people who don't want to buy a machine with XP loaded and then have to reload the 64bit O/S. I'm one of those people.
Well, as you say it is very difficult to determine just what WIP is made up of. Frys will occasionally have a Duron and MOBO package for sale, but I've just assumed these were returns from the channel that never got sold. AMD has been rather specific in its' guidance about inventories, saying that there was not the sort of bubble that INTC had/has.
According to the q3 10k WIP went up to $534M from $504 in q403 and total inventories increased from $697m to $807m during the same period. Finished goods did go up from $149m to 212m during that period. Since we're talking sellable units we need to talk about finished goods. I have no idea what the lower of (cost/market value) of the dies in inventory is, but will assume an average cost of $25. At that cost q3 finished goods represents about a 2.5m unit increase over q403 or about 8.4m total units. That's not that much, about 1 quarter, and doesn't seem to indicate much of a buildup of inventory. Besides inventory goes stale rather quickly so keeping things around like Durons is probably not the norm.
As I see it the crunch is going to start showing up in q2 with the release of the MS 64bit O/S. Demand will be for AMD64 products. How much production is shifted to AMD64 products will determine how fast existing inventory is drawn down. The halo effect is going to affect all AMD products, but it's hard to say what kind of demand there will be for say Duron products. Heck there may be k62 products still in inventory, who knows.
Anyway AMD isn't going to sell every processor in inventory under any circumstances, there will always be some that can't be sold and those that haven't been sold through channels yet. Most of the stuff listed in q3 finished goods was probably in the channels on consignment. Even in a period of heavy turnover there are a lot of finished goods sitting in the channels. This is like the water in a pipe needed to prime the pump. That amount isn't going to change much under any circumstances.
On top of it all one of the pictures I saw seemed to show INTC venting its' heat into the AMD machine. Talk about handicaping.
Very, very strange day. I wonder how many of the shorts closed out their positions sub $21? Of course it hasn't hurt that the general market has gotten stronger as the day has wore on, but if AMD goes positive that would be a very big change in the markets thinking.
Today seems to have started with the continued short selling, probably some left over call selling, a Pru downgrade, and finally some nasty flash news. With some bats ears stired in and the removel of a lot of stops the recipie was complete.
I wonder how many fund managers were on the horn with the office this morning from their remote retreats? Anyway, right now over 20m shares have traded and AMD is down $.25. This reversal is more of a moral victory, a vindication of AMD's fundamentals as I see it. It's still going to take earnings to get this puppy moving again, but we may have defined a new trading range today.
It is amazing how most of the analysts seem to conviently forget what happens to profits in a fixed cost business when it runs its' production facalities flat out. Not to mention the room it provides to raise ASPs. It is too bad AMD probably won't be able to take much market share from INTC, but we stockholders will be crying all the way to the bank.
I really want to know what percentage AMD64 is of total q4 production(60%?). With AMD no longer being a one trick pony, there's a much larger line of products to support. It's going to be difficult for AMD to shift too much production to AMD64 products without doing damage to to the other markets it has fought to create. Tough decisions will soon be looming for AMD management.
Stops being taken out is probably part of it. 15M shares traded so far, not bad for the week before Christmas. Well it looks like the worst is over, maybe someone figured out that flash really isn't the biggest part of AMD's business? Anyway, it's probably the little flash guys that are eating it worst. AMD's continued move to mirrorbit and 110nm are probably insulating AMD from the worst of the priceing action. AMD's probably not liking the market but doing ok. Not much of a surprise to us that this is going to be a CPG quarter, but without flash the q probably isn't going to be as big as a few have speculated/wished.
Mad rush to the door. Who yelled fire?
From a strategic point of view today is a very good day for the bears to mount an attack. With the flash news and a lot of the big boys already on vacation AMD's price is subject to any strong gust. The bears are doing very well with the repositioning they did at about $25 so any change in the wind will probably result in a lot of covering and a big reversal day. Still, that day is not going to be today or tomorrow. Probably after the first of the year when AMD does the sort of forecasting it did last quarter.
Your right pre-tax loss of NT$862 million for the quarter. I misread the NT$150 million as being the quarterly figure. That does include a currency loss of NT$170.
Since both AMD and Macronix are Saifun customers I wonder how much of the parternering speculation is being driven by mutual needs. Micronix being on 250nm and moving to 150 seems to indicate that they are way behind AMD and INTC in flash tech/production capabilities. Still the mention of of an idle 300mm fab is interesting.
Personally, I think Hector believes that AMD has the design and cost structure to go head to head with INTC in a flash price war. This Macronix talk may be part of a plan to flood the market with NOR flash that only AMD can make a profit on.
In any case if AMD is serious about ORNAND they will need a lot more flash capacity.
This is probably what's behind the price pressure we saw yesterday and this morning. Still Macronix is projecting a profit. AMD with its mirrored-bit making up more of the product line probably has a slightly better cost structure. Still, while not expecting to see a loss in flash for AMD, BE is beginning to look pretty good. Probably somewhere between BE and last q's $15m.
Given the relative contributions of CPG and flash to AMD's bottom line it looks like the market is probably over stating flashes importance.
It's always nice to have a little powder left for moments like this, but we're probably not as low as we'll go yet.
Yeah, it's hard to say how this whole thing is going to work out. There's going to be a lot of merging/new alliances going on as markets consolidate/merge.
One of AMD's major strengths has been its' lack of interest in the chipset/GPU business, this has fostered all kinds of 2nd/3rd party innovations that will only grow next year as INTC follows NIH to the tee. But as the market continues to shift towards platforms there's going to be more of a need to choose partners. Choosing well for AMD is going to be critical. At this point I'm not sure who would make the better chipset/GPU partner, or if AMD even needs to get married just yet, but NVDA has to be in the running. By the end of next year pressures should start to build as FAb36 gets near to coming online. Fab36 coming online is going to radically alter the current AMD.
I don't think it's worth the trouble, besides I keep a very clean machine, constantly purging old Internet data as well as running add data checks.
With the number of sites I visit daily keeping the machine clean is a must or things soon start to bog down.
Sorry, I read so many articles I've already forgotten what else was in it.
Yeah, as all these various forces start to meld we'll be seeing GPU manufactures start to poach on the processor business and vice versa. It's only a matter of time before some processor cards start including GPU functionality.
A lot more differentiation is ahead as the PC, as we currently know it, is relegated to the business/scientific markets and the Home PC drifts towards a more communications/storage like device. And then there's the notepad and smaller market where low power and functionality will be competing.
Lots of changes ahead.
I think the too large had to do with too many transistors in the Prescott, something that has baffled more than one reviewer. No one seems to know what all the extra transistors are doing. Just my speculation.
Sorry about the posting errors, but as you know each site has its' own standards. Personally, I think it's better to over post than under which is why I don't get disturbed when the same thing is posted more than once. A lot of times the first article gets revised.
Sorry about that. Try this about INTC's inventory.
http://www.investors.com/editorial/tech01.asp?v=12/17
INTC inventory problems
http://www.computerweekly.com/articles/article.asp?liArticleID=135918&liArticleTypeID=21&liC...
Them...
Whittington says the company's older "Prescott" Pentium 4s are going unsold because they are too big and generate a lot of heat.
Intel "has some stuff nobody wants," he said. "I'm surprised the accountants aren't insisting on a one-fell-swoop write-off."
Me...
from the Saxman
Infineon creates 20 nanometre flash cell
http://www.theinquirer.net/?article=20253
Wow, wild and wooly out there. This doesn't look like anything AMD specific(SOXX down 1.7%, AMD down 1.06%). The low volume is amplifying the situation, which isn't probably being helped by triple witching. Could be just some of the big boys cleaning up before taking off for the Holidays.
A better question is whether this is the best INTC can do? I don't doubt that INTC could come up with an on processor MC, but to do so would mean giving up some control of the chipset. Given the low-latency advantages of having the MC on the processor it means that INTC has chosen higher profits from chipset sales over creating a more competitive processor.
In this case it's not another one of those laws of unintended consequences things, but a calculated plan that maintains control over proprietary chipset standards. It's interesting that INTC hasn't yet learned that this sort of thinking is at the root of its' problems.
You know it was only a couple of years ago that chipset makers wouldn't make AMD chipsets out of the fear of INTC retaliation. Now we have virtually all the chipset companies falling all over themselves to produce AMD chipsets. One might ask oneself just why that is so despite the "INTC inside" program? Could it be that INTC is forcing them into AMD's arms through its’ use of licensing fees and processors that underwhelm?
Well despite the low volume someone has taken a shine to AMD. This kind of action is one reason why I continue to think it's more risky to be out of AMD than being in. What I'm really worried about is being out when something like a DELL announcement happens. Not that something from DELL is likely, but there are so many other things that could move the stock it just looks risky being out. Timing the rise/falls is the risky part now not the direction.
Yes I remember MTH, I'm sure INTC has their reasons.
I have no idea what INTC's plans are, better ask Elmer. I know why INTC doesn't want to put the MC on the processor(control chipsets) but sooner or later I expect INTC to alter course again and follow AMD's lead.