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.
i don't think anyone can successfully compete with intel for the long term given the apparent capability of intel to fill all niches.
it wasn't always the case that intel was intent on sopping up every x86 opportunity. i think that has passed now that they're pushing hard on HPC and Silverthorne ends of the spectrum at the same time.
that doesn't preclude someone trying however and making life difficult for intel in a particular area.
and the market seems capable of putting down large on longshots.
gb
i love it!!
gb
use ibm as a patent laundry for one.
gb
i think amd will go bankrupt by '09 if they not bought.
i fully expect them to throw a legal tantrum about what they willingly signed as a cross license and try for a stay of any attempt to enforce the terms of that agreement.
remember it is intel that will have to assert patent infringement claims against amd upon default.
gb
i think intel will never lack a competitor. the size of the market opportunity will be too big for investment dollars to ignore.
there may no longer be a company named amd however.
there are lots of issues for another company to overcome including patents and fabs.
doesn't mean no one will try.
gb
anyone have any idea what this is about?
is there precedent for such a meeting?
gb
now this is truely weird!
henri who is always crowing is now volunteering info about mss losses in servers?
is this amd's idea of how to warn for q2?
???
gb
maybe it just seems like all day...
run XP all day long
so who could the shareholders sue? can't get blood from a turnip.
the tangible net worth of amd is around $5/sh and even that is pretty entangled in debt covenants.
i suppose they could try to individually sue the officers like hector, dirk and henri.
to what end? dividing their total assets among the various shareholders would just line the pockets of the institutions who hold 80% of the company already.
and i doubt that would amount to much more than pocket lint to them.
i think the only likely course is a slow and bumpy ride to single digits based on continuing quarterly losses throughout '07 and '08.
and a "new business model" after that imposed by whoever gets control of the company by '09.
note that most of those who have tried "new business models" haven't created a viable company rather lined their own pockets while giving the appearance of a "good college try."
gb
agreed on the legal requirement.
however when rolling up revenue numbers it often becomes pretty obvious to many in the sales organization if there is a significant miss.
since it is too difficult to control leaks companies warn to avoid regfd issues.
my reason for posing the question is that it must be pretty obvious to many inside amd that they aren't going to meet their guidance. especially with announcing a price cut which will take effect in the first day of q3.
that will either mean customers will demand a pullin of q3 prices or push out their orders. oems will likely get the q3 pricing, the channel will likely just pushout since they have little pricing leverage through distribution. either means an additional hit to revenue.
hardly any risk to a pushout when they can just pump their intel skus during a traditionally slow quarter.
gb
thanks for the info. this should be fun to watch. at 283mm2 perhaps the hpc guys will take all of the barc output assuming a useful clock rate.
i wonder how many dreadfully slow ones will get produced in pursuit of saleable speeds?
gb
ok, that's one. and considering it's a one-off (or nearly so) it isn't surprising that it was done.
but does that really represent a serious oppty for a company like amd to absorb early output?
gb
isn't it approaching "warning season" for amd?
iirc, they estimated q2 to be flat to up.
doesn't seem likely given the limited data has been on the news wires so far this q on pc volumes and intel/amd price moves.
gb
when the p4 tapped out on prescott even at 65nm it required a major architecture redesign to get things in balance with what the process could deliver, i.e. core and core2.
i think the problem that amd is having is a combination of a process wall (lack of gate oxide scaling) and insufficient resources to staff a real architecture redesign.
barcelona may well be their prescott in terms of not meeting expectations.
gb
i always shake my head when i read about "upgradability" as an enduser or IT concern. i'm sure upgradability is important to oems who have platform they'd like to refresh. i'm also sure that hotrodders like to slather heatsink grease all over all sorts of contraptions.
but the vast majority of folks who own a pc never even open it to vacuum out the dust bunnies, seldom add anything internal and certainly don't replace processors.
do many IT shops actually take a server off line and start removing and replacing processors and heatsinks? somehow i doubt it.
iirc, my infrequent conversations on this with IT folk at intel, servers were put online for a purpose and when they broke (years later) they were retired.
gb
interesting, but it only makes me wonder how intel is able to ramp 65nm and soon 45nm so much faster than the other members.
perhaps the sematech stuff is merely a starting point rather than a manufacturable recipe.
it will be interesting to see if anyone other than intel can do hi-k, metal gates in '08, or even '09.
gb
some larrabee info:
http://bt.pa.msu.edu/TM/BocaRaton2006/talks/davis.pdf
gb
wbmw: i think you're right on target.
amd doesn't get their own physics.
if they don't get a recipe for thinning gate oxide that works on both p and n channel devices they won't get much if any benefit.
each time intel demos their 45nm based systems it must make ibm and amd crazy.
intel hasn't disclosed their metal gate formulation yet. it seems like they're trying to keep it a trade secret rather than patent it.
it's going to be really interesting by the end of '08 as intel is producing everything from octals to silverthorns on 45nm with their "too many fabs" while amd is still explaining why lack of 45nm is "more cost effective."
fun to watch.
gb
since nehalem is 2h08 for production, i'm guessing that it must be nearing tapeout.
i wonder if we'll see any nehalem demos at fall idf or if they will wait until fall analyst day?
gb
opteron 8way? lemme guess
does the opteron support more than 4way natively? if not it takes a custom chipset.
developing and validating one is expensive.
who would do it for a chip that only represents 20% of the server market?
oh, and it has to offer a performance increase much more than two four way servers.
of course there's horus...uh, wait a minute...
gb
octal cores in that time frame won't be a battlegound...it'll be a walkover for intel.
gb
and they have a lot of plans for "new things".
silverthorn tiny but lots of them. larrabee.
i very much get the impression that intel has figured out that they have a unique advantage and are going to take maximum advantage of it while they can.
that spells trouble for the "unaligned"...
gb
shouldn't duke be putting the lyrics to "hotel california" up by now?
you can checkout anytime you want, but you can never leave...
gb
per weird al yankovic:
it's weasel stomping day!
http://www.azlyrics.com/lyrics/weirdalyankovic/weaselstompingday.html
gb (must be heard to be appreciated...)
http://www.intel.com/pressroom/archive/releases/20070522corp.htm?iid=pr1_releasepri_20070522r
intel is selling their phase change to this as well.
still makes me wonder about the micron jv however.
intel will have ~50% of two companies competing on nand.
sorta weird...
gb
i wonder what happens to intel's flash fabs? weren't some of them collocated with other intel fabs?
gb
actually richard wirt was scooping up software folk around the time of the 386 (1986) launch including compiler folk from the various failing alternative architectures of the time.
gb
I think there are several complicating factors here. Vista no doubt caused problems for all three graphics vendors. Separately DX10 is/was a challenge. Intel's change of underlying hardware was another big change. Lastly Intel's increased focus on video performance to provide graphics benefits to those outside the gaming community provided increased stress.
That's four stressers at once and it will take some time to get the resources up to the point where they get on top of the problems.
Intel has good software folks. It likely has too few applied to graphics since it wasn't as important as it has become now.
I expect that to change. It should be interesting to watch especially as Larrabee gets more visible.
However in the short term it must be pretty frantic inside trying to get G965 and the 3X chips up to platform expectations.
gb
quite good...if you judge by the tattered condition of their footwear!
gb
before retiring i often sat through chipset planning meetings. one of the hot topics is when/how to discontinue support for legacy i/o due to the new process not having the capability for dealing with 5V for instance.
so it's no surprise that as chipsets inherit cpu processes there are tradeoffs to be made including external fabs on occasion.
gb
all of which says to me that barc is all "bark and no bite."
if intel can wax it (and i fully expect them to) that easily then amd is toast...and they know it.
gb
i cracked up listening to hector's and dirk's response to a shareholder's question about why they aren't painting the world with barc info.
the answer was that if barc info was public, powerful, fast-follower intel could come up with a competitive response too quickly.
any shareholder who believes that crap deserves what they get.
if amd really has something unique it would take a product cycle for intel to respond. if it isn't unique and merely some "innovative" benchmarking they're toast anyway.
pathetic...
gb
PSO was asked about Larabee at the end of the Q&A and the questioner suggested that Larabee was just a highend niche product and didn't compete with NVidia.
PSO set him straight on that and it's very clear from his response that Intel intends a pincer movement on the graphics market: continuing to ramp integrated graphics while attacking and pushing down from the top.
Graphics is gonna get very ugly in the Larabee timeframe.
gb
I finished listening to the SAM call last night. Near the end of the Q&A session Joanne Feeney asked a few questions about AMD's asset lite strategy. PSO and Sewell (legal) responded well I thought but Feeney wouldn't take the answer and was finally cut off.
The rest of the analysts and I got the message: there is definitely a contractual limit to AMD's ability to outsource.
gb
and how about those absolutely sleazy black-scholes model shootouts on 45nm quadcores?!!
gb
PSO's presentation today was all done on a 45nm Penryn DP desktop processor. Of course we don't know what speed it was at but it definitely shows confidence.
gb
i'm enjoying watching intel closing out all the niches that amd might chose to exploit on three levels: shutting out amd, expanding the tam intel can exploit and giving me something else to spend my money on!
gb
not a rude reference to Hector or Richard
now that was funny!
gb
Mercury says Intel @81% in Q1. Further share loss predicted for AMD.
http://blogs.barrons.com/techtraderdaily/2007/04/24/intel-new-data-shows-share-gains-in-desktop-note...
gb
i was thinking of the economy in general.
i'm still specifically negative on amd but not invested in amd or intel right now.
as a retired intel guy i still root for the "home team" and many of my still employed friends.
gb