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.
keith
"The street" represents institutional capital. Mostly East Coast bankers, considering other things than just current outlook for buying or selling stock for their clients and for themselves as well.
Apart from that, the fact Hans Mosesmann (Soundview) is a former AMD-employee sheds a light on the message as well (although for me only insofar as i have no doubt about the credibility of some internal affairs indeed going on wrt board). K.
Keith
I´d personally be happy if they finally canned Jerry
Dont you think that answers your question already? Put it this way: JS is probably worth a billion USD (negative shareholders value currently. If only for his "Chance of your lifetime"-statement May 2002 at ML. As long as he presides the board, the street will be suspective about AMD - rightfully imo.)
Apart from the negative perception of his pay and expenses - which is basically peanuts but the public perception of it is not.
K.
paul
Nothing new, it is a quite complete and correct description of K8 architecture - presented in such a boaring way that i would expect not many make it through the details but most will jump to the conclusion instead.
They basically conclude thats all fine, but except for a niche of mobile workstations nobody would need all that for mobile computing. Finally they hyphenate power consumption and compare it to Centrino. K.
Petz
http://www.siliconinvestor.com/stocktalk/msg.gsp?msgid=19709713
is very very interesting. I basically like the approach.
Im not finished yet thinking about it. Some comments on your numbers: There will be not much taxes for quite a while to pay for AMD. And AMD has driven down costs significantly. That would somewhat change the underlying revenues for future earnings.
And on the approach: Your considerations about stockprice-peak and earnings-peak in 2000 have to be put into the context of brisk change of market sentiment during Q3/2000, and (which is the other side of the very same medal) points to higher expectations for AMDs future earnings at the stockprice peak.
Basically, the historical viewpoint you take is a rear-mirror view. Which is fine. Im just not sure if any projections therof into the future are valid. Maybe looking at 2000 from hypothetical viewpoints within the first half of 2000 would be more appropriate?
I would agree the markets currently look ahead some six months (which historically is a very low horizon). Specifically for AMD, that may well persist due to concerns of manufacturing execution but could increase a whole lot as soon as the street has confidence Fab 30s migration to 90nm will be (more) successful (than Intels). As soon as this would happen, the horizon would expand to 2005 earnings.
K.
mas
No problem. Grantsdale can deal with DDR as well.
The real benefits of the platforms are not affected at all:
1. Sufficient power to allow scaling Prescott.
2. LGA is a cheaper solution, leveraging 90nm/300mm economies.
I dont think Intel will be cheeky enough to put too much force behind DDR-2. RDRAM is not forgotten yet.
Although, there are many many people out there going just for the latest product at any price even if it is inferior to its predecessor. I mean they sold a whole lot of Willamettes, so why not DDR-2-boards as well?
Now, it will be interesting to see how many Grantsdale DDR and how many Grantsdale DDR-2 boards will show up. That will give us an idea of what Intels marketing-research data tell them about the education-level of buyers.
My gut is inclined to believe the "most powerful processor" combined with "next generation DDR-2" could well be embraced.
K.
elmer
Guess that is pretty clear in dimension from previous postings of this thread. AMDs manufacturing-ratios of K7 and K8 will liley be adjusted to market needs every week this year, so any specific number would rather be a market-based than a manufacturing based forecast.
K.
wmbw
Ok. I have a high level of conviction your plan will work out.
K.
paul
That is for the New AMD.
Quite unusual for those following AMD for a while before, we learned to look at these blocks as mobile objects, tending to slide once and then. I mean the only predictable thing was the direction of the moves...
K.
chipguy
Well, following your thoughts, the focus currently lies on server, then mobile, then desktop. Under these considerations the huge die is no problem unless you want to penetrate valuespace with it. Then, it depends what numbers they can make in mobile segment with 130nm K8. Every indication points to an upcoming low-voltage CPU for Thin and Lights - my guess would be sometime in summer.
Then, in summary, if you add server plus FX plus mobile plus desktop performance, 6 M desktop CPU make still a slow ramp, but a deliberately chosen strategy to migrate under conservative assumptions (e.g. no windows64 this year) - which is admittedly a paradigm change to what we would have seen in Jerry's times. Which would have been a lot more entertaining, but not necessarily more successful (carefully said ).
What I have seen from the "New AMD", better said how I read Hectors handwriting, 6M desktop CPUs seems to be in the right ballpark for me. Probably accidently, I might add, looking at the source it comes from.
K.
sgolds
Yes I believe AMDs roadmap is indeed their best guess (although a tad on the optimistic side: It has a second generation 90nm process for end of 2005 on its map. (I read: 2006)
Now, a process node has to run at least two years to make money. 2006 + 2 = 2008. Sounds right to me, until then I would expect the transistor-geometries ready for volume production I assume are required to make 65nm manufacturable.
Anything else I'd like to see before I believe. It is possible anytime that you find gate oxides or modifications of current geometries to make it manufaturable earlier - but that i would consider a lucky strike which is nothing you can really predict, let alone put in on roadmap.
K.
wbmw
Thanks. Not that I would intend to fill elmers temporary vacancy here, but your number exceeds by far what my model says Fab30 will be able to churn out this year for K8. Plus, i dont see the market absorbing six million AMD-CPUs per quarter at reasonable ASP which they need for the fairly huge die they use in the current node anytime this year.
Later, in a Windows64 and mature 90nm process environment, I can imagine it. But both of it is next year imo.
K.
ps: Given your expectations, did you bet the farm on AMD?
dougSF30
As i said, i am waiting for a second source to confirm.
Meanwhile
http://pc.watch.impress.co.jp/docs/2003/1216/kaigai051.htm
and
http://pc.watch.impress.co.jp/docs/2004/0116/ubiq42.htm
(thanks jsw for translating it to german on w:o)
basically says C1 will come in Q1 but yields very limited volume of higher bins, only D stepping and LGA775 in Q2 expected to yield higher clockspeeds in volume. Samples of D stepping not yet been spotted in the wild.
K.
sgolds
Look what AMD has on its roadmap for Fab 36.
Intel will make Tejas using ultradense SRAM which is possible in 65nm and call it a 65nm CPU.
As for the asian foundries, I see Chartered to be supported by IBM (as soon as it filled its own foundry) and UMC supported by AMD (as long as AMD does not come up with the idea of offering fabspace in FAB36 and ramp it up to 20K WSW next year). TSMC, I dont know.
K.
sgolds
When do you see 65nm? K.
mas
http://80.237.203.42/vbulletin/showthread.php?s=d0891fd7cacb5b4ad5fc5cb0745fb2e8&threadid=119090
Too early to comment on. Waiting for a second source to confirm.
K.
wbmw
This number far too low or far too high in your opinion?
K.
wmbw
Well, currently there is only foundry able to deliver 90nm afaik.
I dont believe any of the asian foundries has a 90nm process ready for volume this year.
As for the article and the above, i understand the statements of these guys just as responding to IBM's Meyersons comments earlier last week.
As for (PD)SoI, i am inclined to believe it still helps in the 90nm node, if only for some part of the leakage issues; but I wont argue about where exactly Intel would be "disadvantaged" (nice word btw) or even handicapped in their 90nm process. It could be many other things, attributing anything only to the lack of an SoI-Design would not be only premature but most probably incomplete at least (as every unicausal explanation would be).
To keep it simple, I look at what comes out and when from their fabs and compare it to what others have to show: Two 90nm products on the market within weeks from now will allow a comparison. I sure expect a lot of bragging from IBM aboout the specifications of these products - these guys have a foundry to fill.
And it will be interesting to watch the street's perception of it.
K.
sgolds
Right on the spot. As easy as that. K.
Edit: The "role reversal" happened already long ago (launch of P4 Willamette) imo.
mas - thanks
So then no Spec-boost. As for 512--->1MB, same here.
K
joe
I am not good in estimating impact of RAM size on SPEC, TPC and the like.
In real life, I know for sure every HD-access hurts badly when it comes to performance.
K.
RGood
Well then.. All the best and good luck. Courageous to bet against the CFO of the company with its stock. I am impressed.
K.
RGood
Hmm.
Well. I guess you are allright with a longterm Intel-investment. So if thats your investment style, I dont see any reason to change anything - completely in line with the Analysts you have faith in.
Short term, you might consider that Andy Bryant (which you would agree is as conservative as Craig) sold the majority of his own Intel-position a while ago. And you might consider that common consensus is that we are right at the beginning of a tech-growth-cycle, which is usually the time for an upgrade of tech-stocks.
K.
p.s: ...institutional driven investors....
You put that one nicely. Now, think about what it really means.
Chipguy
Thanks lot. Looking forward for SPEC submissions then.
Doubling RAM should be able to do something for Scores.
(As long as all timings are like with 1GB-Modules, that is.)
K.
keith - "AMD´s" PC3200-Reg standard
Still no JEDEC-spec for it? Probably never will be?
Anyone knows SPEC regulations allow use of non Jedec-spec'ed parts for submissions?
K.
One might expect some capital travelling from Santa Clara to Sunnyvale for hedging purposes.
One percent of Intels Capital would do wonderful things for the AMD common. K.
mas
What I heard is that most of their 2004 capital spending is related to 65nm. K.
mas
90nm looks like it's going to sort out the men from the boys.
At least that's definitely like Bernard Meyerson wants it to be perceived. For me it rather sounds like he actually intends to leverage IBMs bragging rights somewhat for the launch of 90nm-G5 - which is very understandable, IBM still has a foundry to fill. Basically, 90nm challenges are known for long, and the pattern of increasing challenges from node to node is not exactly new either.
Thats all fine with me - it should help for a better perception that (PD)SoI is still very helpful in the 90nm node.
Which imo will be a way longer node than most believe today; from current viewpoint, 65nm is beyond any horizon.
(Let alone 65nm volumes next year )
K.
doug
I dont see a pattern like first half of 2000. Looking at the beginning of the cycles and comparing K8 to K7, this time it has started a lot slower, consistent with market growth-rates, which for last quarter were slower than increased capacitity utilization and capacity-growth. Entering two slower quarters now there is good reason to extrapolate this trend, implying competitive pressure as well. In essence, I think this cycle will not be like predicted fast and steep (bubble-shaped), but slower and therefore persisting for at least two, probably three years. I guess the technological barriers ahead support a pattern like this as well: Everybody in this industry will run into the 90nm-challenges within this timeframe, preventing the capacities to overboard. For today, look at Intel, they clearly overestimated the growth for last quarter. Look at DRAM, Flash. Same pattern. No Supernova, no big bubble ahead. Healthy growth in a competitive environment instead.
Good for the industry, good for consumers. AMD common is not in danger to be damaged by any moon-collision. (Sigh, i would not mind at all if it was) Just headed for a fair value of 20 times 2005 earnings within the course of this year. Fourty bucks, maybe fifty if the growthrate will be awarded with a 25 multiplier.
Which leads exactly to the 2000 target, just later in the year. Unless the market anticipates all that and we indeed hit the 50 Bucks already next week. I wouldn't mind at all.
K.
ps: Now, I could have made the long story as short as: Peak in April-June timeframe is easily possible. As well as earlier, as well as later. I mean, as long as I look at the markets (which is admittedly only some twenty years), they never behaved like I thought it would be economically reasonably for long.
RGood
Well...
Now, to be a lil bit more precise wrt "follow the money": You might agree the analyst side is unusually quit after an Intel quarterly report, do you? Next, there is a hell of a lot of money from their employers in Intel stock, right? (And yes, I know analysts are independent from that ). Then, look which stock has huge upside volume in the sector.
I dont know how much of this is to be understood as hedge-moves of institutional capital. But it could be worth a thought for your long position as well. Unless you are convinced Intels money-machine is unbreakable and intend to hold the position longterm. In this case, no action is required.
K.
OT Buggi
w:o seems to be broken.
Just to confirm, I take the bottle bet. (I see FASL over +18% as well, so would not be happy to actually win it)
Sort of hedge , to put it that way.
K.
RGood
I'm not sure I understand, which stockmarkets are you looking at?
As for the US-markets, my screen shows mostly green figures...
And I'm not sure I understand which CC you were listening to.
As for the cheerleaders... (yawn)
Now, to get an idea of what is going on here, i would suggest to follow the money
K.
keith
Dilution from Conversions of
Senior Convertible Notes (400M/7,37)
and from Convertible Bonds (500M/24,??)
K.
edit: Thats for the # of shares. The profitability levels you cannot calculate easily. Has been given out by AMD IR, the figures are from the top of my head.
@keith
Depends on the level of profitability...
Below about 7 cents it is just short of 347 M
Above that, below 23 cents it is around 400 M
Above that, it is around 420 M
K.
Joe
It currently does. Next generation CPU will be fabbed by IBM.
http://portfolios.abcnews.go.com/quotes?tid=&tab=news&headline=no&source=reuters&pat....
K.
Joe
Except, the roadmaps I have seen don't have any large cache (processors with L3) Prescotts at all. Not until Tejas core.
Yes.
Looking at the roadmap I see Intel focussing on costs for a while, playing its strenghts. Bottomline, if we cannot outperform them this year, we will outproduce them.
Economically reasonable without any doubt and completely acceptable if it is meant as an interim-strategy: Even if they allow ASP to decline some 20% this year, they can make a ton of money as long as they can make the stuff 25% cheaper due to 300mm/90nm economies. (Besides they have a strong product in the fastest growing mobilesegment.)
I expect this to be Andy's message tonight: Look at gross margins and earnings rather than revenues going forward.
Personally I am not sure it will play out good for the Intel Common short term; an average investor will hardly be capable to look beyond guidance for declining revenues, but well, maybe the bunch of cheerleaders (the ones asking questions tonight) will help to get it down the throats.
Now, for AMD common I frankly have no clue what it means short term. Probably nothing good - but you never know.
K.
EDIT: On a second thought, Intel could change guidance-policy and only give earnings - numbers starting today.
OT sgolds
Lol. Sure. Thanks for Relaying Credits
Keith
I might add I dont expect the specific date to be made public in advance.
But i would assume they do it asap and let us know the day after.
K.
OT sgolds
Thanks for looking over it and your comment. K.
keith
To be determined yet. Jan 29 the earliest possible date afaik.
K.