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.
Re: INTC currently gets about 25% of their revenue from something OTHER than CPU's
They have a desktop/servers group and a mobility (notebooks/PDAs) group. Each is dominated by PC CPU sales plus tie in sales of chipsets and motherboards for those CPUs.
The "other" group lost $464 Million on $47 million in sales so expanding that might not be Intel's best stategy....
Three Months Ended
October 1, September 25,
(In Millions) 2005.00 2004.00
Net revenue
Digital Enterprise Group
Microprocessor revenue $ 4,936 $ 4,520
Chipset, motherboard and other revenue 1,434 1,346
6,370 5,866
Mobility Group
Microprocessor revenue 2,331 1,571
Flash memory products 572 638
Chipset, motherboard and other revenue 640 352
3,543 2,561
All other 47 44
Total net revenue $ 9,960 $ 8,471
Operating income (loss)
Digital Enterprise Group $ 2,162 $ 1,808
Mobility Group 1,402 796
All other (464) (231)
Total operating income $ 3,100 $ 2,373
Re: Before AMD has AVIS status I think Intel would fight tooth and nail on prices - and once their products have true parity again this tactic would hurt AMD.
I think that Intel fighting tooth and nail on prices would hurt AMD even if there is no product parity.
Another factor facing both companies is that the CPU isn't the contraint very often, these days. A mid range CPU is more than enough for almost all users.
But both companies can still hope to make some money, and AMD can hope for significant increases in revenue market share (it's mathematically impossible for Intel to gain more than about 10% revenue market share, ever).
For Intel, I just don't see much upside at this point. As something of a metaphor, how much upside do you see for residential builders, going forward? They've had a great run for a number of years, but for most investors, the expectation now is that they'll either be doing somewhat less well, or a lot less well (possibly disastrously poorly) going forward. There just isn't anywhere else for them to go, for a while.
If Intel decides to go to a scorched earth pricing policy, it will definitely hurt AMD, and badly, regardless of whether AMD has somewhat more desirable parts or not. Doing that would hurt Intel badly, too.
Let's hope it doesn't come to that.
Re: The question was: how would you grow intel's business..
I don't think they can. It's like GM a couple of years ago - they have nowhere to go but down
Their best hope is stagnation. They still have close to 90% revenue share - virtually no room to grow there. And they also have a competitor that has better products and is growing quickly.
They need a miracle chip that will give them back their monopoly pricing power, and they knew this 5 years ago and they came up with IA64 as the way in which they could grow their business - which, had it worked, would have given them copyright ownership of the computer hardware business. Intel's strategy was that the inevitable move to 64-bit computing would bring along with it an Intel only hardware API.
But IA64 failed and Intel has had to scramble to offer AMD64 clones.
Intel's business plan is broken.
All Intel can do now is to try to slow down the rate at which this becomes a Hertz/Avis business versus its current Microsoft/Everyone Else business.
They tried competing in new markets (remember when everyone was claiming Intel's new LCOS TV's were going to "take over the market" this year). Intel isn't really any good at anything except making CPUs to sell into a market in which they don't face any competition.
Maybe Intel could open up a restaurant chain or try to out do Wayne Huzinga selling used cars.
Re: If you were in charge of Intel, how would you run the show instead?
I'd have them stop engaging in criminal activity before lawsuits and the justice department wipe them out altogether.
If you were in charge of the mafia how would you have them handle their business?
Re: since Intel had no criminal tactics
They did. I've spoken to OEMs who described Intel's harangings regarding any use of AMD parts.
AMD's original Opteron launch was badly damaged and seriously delayed for many months after Intel's extortion of Solectron left a big hole in the Opteron platform.
Looking at how well Opteron has done, imagine if much of its platform hadn't been delayed for 6 months. Using criminal tactics to steal 6 months of a lead in the tech industry is a terrible crime.
Intel did similar things during the original Athlon launch.
A fair punishment would be if the courts made Intel warehouse all of their new parts for 6 months before allowing them to be brought to market. Say for the next 5 years. You thought actions like that were find and dandy when Intel was doing it to AMD, how about if the courts award Intel a little of its own medicine?
So Yonahs produced in Janary couldn't be sold until July. Meroms produced in October would have to be held in warehouses until April of 2007, etc. If that results in shortages of parts for OEMs, they just have to accept that as something that comes with their use of Intel.
Let's see how well Intel's products compete with an artificial 6 month delay.
Re: For $2,500,000,000 Intel will get the new logo..no dropped "e", no "inside" and a circular 'dual' swoosh thingy.
Looks very similar to old logo.
Re-arranging the deck chairs on the Titanic comes to mind...
Maybe Intel can hire Celine Dion to write and sing their new theme song!
Song: My Chip Will Go On (Itanic) Lyrics
Every night in my dreams
I see you. I feel you.
That is how I know you go on.
Far across the distance
And spaces between us
You have come to show you go on.
Near, far, wherever you are
I believe that the chip does go on
Once more you open the door
And you're here in my PC
And my chip will go on and on
Love can touch us one time
And last for a lifetime
And never go till we're one
Love was when I loved you
One true time I hold to
In my life we'll always go on
Near, far, wherever you are
I believe that the chip does go on
Once more you open the door
And you're here in my PC
And my chip will go on and on
There is some love that will not
go away
You're here, there's nothing I fear,
And I know that my chip will go on
We'll stay forever this way
You are safe in my PC
And my chip will go on and on
Re: Their vendors do, in warehouses they have in the Austin area so they can deliver on schedule, meaning like 10:32am at the shipping dock... so they can be in computers by 2:15pm!
They love to tell that story, but the reality is that, with the exception of low end Celerons and 15" monitors, Dell's shipping times are more consistent with Dell requesting bids for some of the parts at the time of the order, rather than any of them existing anywhere, much less in nearby warehouses.
That's a great business plan, unless it starts annoying your customers.
Re: AMD Opteron against Intel Xeon
Wow, that was just painful.
Maybe they should be comparing Xeon to Via instead of AMD, it would be a lot less embarassing for the Xeon "workstations."
Re: are you just daydreaming ??
It's the main point of the article I was responding to and quoted.
Re: Central to the effort will be... a $2.5 billion advertising and marketing blitz
If it's a "blitz", then those costs won't be spread over too many quarters. I'd expect that a chunk of that is redirected annual advertising, and that the increment is no more than $500 million per quarter, for 2 to 3 quarters.
That's still enough to show up as a meaningful incremental expense (and reduction in profits), next year.
Between fighting to hold on to their dominance in the mobile market (which they'll do, no doubt, but at the cost of modestly lower ASPS) and the continuning "trench warfare" in desktops, workstations, and servers, Intel will have to work hard to avoid a very visible profit drop next year.
Re: I think it's likely that it will be somewhere in the upper 2GHz, but perhaps 3+ GHz is possible for their Extreme Edition parts.
So if Conroe debuts at less than 2.8ghz, you will consider it to be a failure?
Re: Another lie, Dan. Fleck has been "short" INTC forever
Hey, Tench, take a look at a 5 year chart of AMD and Intel.
It's a real hoot!
Start here and just add AMD and make it for 5 years:
http://bigcharts.marketwatch.com/intchart/frames/main.asp?symb=intc&time=&freq=&country=...
Re: I am glad to see they are at least trying.
Good point.
Re: so I doubt they will grow there
AMD has been capacity constrained - that ends next quarter. Just how much of the new capacity they'll be able to sell is something we all just have to wait and see.
Meanwhile, Intel is basing its hopes for future growth on the consumer electronics market:
LAS VEGAS–Here at CES, Intel demonstrated their intention to enter the digital TV technology market with a new Liquid Crystal on Silicon (LCoS) chip. Though the technology has been problematic in the past, the company believes it has overcome the necessary challenges and can mass produce low-cost, high-quality LCoS chips for HDTV...
...The biggest advantage to the chip, however, will be price. Intel's manufacturing capacity, together with commoditization of components, should allow for high definition televisions in the 40- to 50-inch range that cost less than $2,000. You won't be able to go buy a TV with "Intel Inside" just yet, though. The chip is just now sampling to manufacturers, will be shipped in quantity during the second half of the year, and may appear in a few televisions just in time for Christmas.
http://www.pcmag.com/article2/0,4149,1433336,00.asp
Unless, of course, Intel's foray into the world of consumer electronics turns out to be a colossal failure - but that won't happen, right?
Re: Personally, I don't see INTC or AMD as being particularly good investments for the next two years.
Well, AMD has a huge increase in capacity coming on line right now, and they're moving into higher ASP segments that they were never in before, so there is some upside potential for AMD. OTOH, the net result may be a pretty intense price war.
It's hard to see much near term upside potential for Intel.
If Yonah is a huge hit then they won't lose the mobile segment - that's about the most optimistic thing I can think of for Intel for the next couple of quarters.
Re: How many AMDroids claimed that 64-bit is "convenient" for laptops?
Many of our programmers have gone to AMD 64 Windows for the increased virtual memory space - the JAVA guys, in particular, can't live without it.
For someone just running Office, either a Celeron or a Sempron is more than enough - those guys need neither 64-bit nor dual core, for now.
Unless they want to be able to demo the 64-bit JAVA apps....
Re: I felt compelled to do some crunching on it.
It might be interesting to the add the current Dothan and Turion pricing to that table.
Re: Fleck's enormous ego as an INTC perma-bear is irrelevant as long as their clients are supposedly making money.
Well, yeah.
Bill Fleckenstein started his "long AMD / short INTC" suggested strategy around 2 years ago.
Take a look at a 2 year comparative chart of AMD and INTC....
Pressler seems to have surprising headroom...
http://www.theinquirer.net/?article=28412
OTOH, the cores throttle according to temperature, so the chips may or may not have been running at the bios settings. Note that such throttling would not have shown up in those bios settings.
It would have been helpful to see the results of some simple benchmarks run at 13x, 15x, and 16x multiplier settings, but at the time he posted the article he hadn't yet managed to install any OS - we don't know if that system was capable of booting at any speed (perhaps he's missing some drivers?)
Interesting is that there was almost no temperature change when he increased the voltage by 0.025 volts and the multiplier from 15 to 16 to get to 4.26ghz.
Default voltage is 1.35, so that second bump would have increased power use by 5% to 6% if the core speed actually increased.
He was running 40 to 50c over ambient before the increase so the temp should have gone up by 2 to 3 degrees absent throttling.
Re: It´s NOWHERE CLOSE passing floating gate.
Do you have any links to reports of relative revenue sales at spansion, floating gate vs. mirrorbit? Or is that an informed judgement from reviewing the various hints that seem to be all they've given for some time.
Re: In this little world it is talked about as an AMD (Spansion) and Intel thing...
And AMD has thrown in the towel and begun the process of selling off its interest in flash.
wbmw, you've been told multiple times by multiple people that the same FAB can be used for chipsets and flash.
Re: Only if you can find a link where Elmer agrees with your point of being able to interchange between chipsets and flash on the same production line
I posted that they diverted lines to flash, not that both ran at the same time on the same lines.
Use your head - if they could have run both at once they'd have been able to correct the chipset shortage immediately. Instead it's taking them several quarters to switch back.
They have FABs that can be used for either, and they converted production capacity at those FABs from chipsets to flash.
And now they've paid the price.
Prices dropped after Intel started dumping:
"Intel has been mounting a remarkable resurgence in 2004 by using aggressive marketing and pricing strategies," iSuppli said.
Intense competition on price reduced the size of the overall market. Only $4.2 billion worth of flash memory was sold in the third quarter, 4 percent less than in the previous quarter. The decline marks the first sequential quarterly decrease since the first quarter of 2003.
http://news.com.com/Intel+climbs+flash+rankings+as+market+dips/2100-1004_3-5455080.html
You have to have been tracking flash prices here: http://www.converge.com/www-webapp/info/2-6_pressroom.jsp to have seen that it was Intel's rapid ramp of low density parts in the 4mb to 16mb sized that did it.
Re: Have you been living under a rock? For the past couple years, Intel has been *gaining* in flash, at the expense of Spansion.
Wrong. AMD flash revenue was hit in Q4 last year and Q1 & 2 this year, after Intel limited chipset capacity to expand low end flash capacity, but that's it. Flash gains were a winner for AMD during most of 2004.
Flash Brightens AMD's Dragging Second Quarter
By Mark Hachman
July 14, 2004
As in the January quarter, AMD's revenue derived mostly from flash memory through its Spansion LLC venture with Fujitsu Ltd. AMD's memory group reported revenues of $673 million compared with the $553 million reported by the Computational Products Group, which houses AMD's microprocessors.
AMD's CPG revenues declined 3 percent from the previous quarter, while AMD's flash business grew 7 percent. But both units showed dramatic improvement compared with a year ago, when AMD began to emerge from the dot-com slump.
"We achieved record Spansion flash-memory sales in the second quarter and more than tripled operating income from $14 million to $45 million," Robert Rivet, AMD's chief financial officer, said in a statement. "Spansion maintained clear leadership in the NOR Flash memory market.
http://www.eweek.com/article2/0,1895,1623695,00.asp
And why don't you try posting some links instead of just making this stuff up and being wrong? And the "living under a rock" style attacks make you look particularly dumb when you're once again wrong.
Your point that lines produce flash or logic is correct, but the FABs make both and lines can be converted back and forth over time (and were). Intel's problem is that it does take time and once they'd converted those lines to flash to attack AMD's former dominance of low density flash it takes a long time to convert lines back to producing chipsets.
Re: the question was why Mirrorbit still only accounts for a minority of Spansion´s sales
What makes you say that? I haven't seen recent sales figures, have you?
Units I expect are still dominated by $1 to $2 floating gate parts (vs. $10 to $20 mirrorbit parts).
They were at "low double digits" at the beginning of the year, but we did get this from that last conference call - mirrorbit sales revenue increased at a 300% annual rate, last quarter, up from a 250% annual rate the previous quarter. It's either passed floating gate or is getting close, up from a small share a year ago. And you know from the density that most of those sales are displacing Intel sales.
Memory Products Group (MPG) sales of $516 million decreased 4 percent from $538 million in the third quarter of 2004 and increased 12 percent from $462 million in the second quarter of 2005. MPG reduced its operating loss to $50 million, from an operating loss of $90 million in the second quarter of 2005. The revenue increase over the second quarter of 2005 was driven by record unit sales, increased purchases by the largest global wireless OEMs, and a richer product mix. MirrorBit Flash sales increased 35 percent from the second quarter of 2005 on improved unit volumes and ASPs.
Re: Anand has a big comparison between dual core Beasley 3.46
It does look like Intel could be getting closer on dual socket system performance. But keep in mind that he's testing an available, production 2.4ghz Opteron system against a future, not yet available Xeon. The Xeons may get faster once they finish tweaking the bios, etc. or may slow down, if the prototype demo system was running with some error checking turned off.
2.6ghz X2s have shown up for sale, but haven't been offically announced by AMD, and AMD has new platforms coming, too.
http://www.excaliberpc.com/AMD_Athlon64_X2_DualCore_FX-60/ADAFX60DAA6CD/partinfo-id-562386.html
Energy use is a real issue. In my experience, the problem isn't one of electricity cost (it's an issue, but not a deal breaker). The real issue is HVAC and UPS capacity in server rooms, particularly since the grid looks to be becoming less reliable going forward, so UPS and standby capacity are being carefully monitored. There can be a domino effect once you start passing the original design loads for the server room. You need more UPS, more HVAC, and more standby power for the systems and HVAC. You then sometimes find that your ducting or chiller pipes are inadequate to support more HVAC and that your building transformers are indadequate to support additional circuits. Once you start replacing transformers and running new supply and return lines through walls to where the chillers are, costs can get pretty impressive. We've been paying high prices for (relatively) low performing EE Opterons for one of our server rooms for just this reason. We'll be accepting that compromise in more systems going forward. And note that 32-bit chips can no longer be taken seriously for servers - The new Microsoft Exchange Server and other server applications aren't even available for a 32-bit OS.
Power could end up being a winning issue for Intel if they can extend the Yonah architecture to 64-bits without the big increase in power use we saw in both Athlon XP and netburst when they were extended to 64-bits. I doubt Intel will be able to do it, but they talk like they can, so maybe they can (note that they talked like they could do it with netburst, and they were wrong). We'll have to wait for Woodcrest to find out. By then AMD will have gone to a new core, a new platform, and added SiGe, so Intel stands a good chance of finding itself remaing fairly far behind but maybe they'll catch up or even move ahead.
Re: So no pointer huh?
LOL !!
Pointer to what? You make something up that I didn't say, then insist I confirm it?
Re: Show us one link where Elmer said this.
If I find posts from Elmer that say so, will you stop posting on the thread?
Re: since Intel has been recovering share in the flash market
Intel has taken share only in the low end of the market and only by dumping well below cost. They dropped the prices of the 4mbit and 8mbit parts by up to 80% - the part of the marke that AMD used to dominate.
It cost them a lot of money to dump those parts below variable cost and it also cost them their reputation as a reliable supplier since they used badly needed chipset capacity to do it. They've also had to go crawling to the chipset makers and give up their former royalty revenues to get some more chipsets available for their CPUs.
But they did gain considerable low end, money losing, flash share in exchange for those losses, and that definitely hurt AMD.
Re: Mirrorbit turned out to be an overhyped technology
It may or may not have been overhyped, but it knocked Intel from clear leader in NOR flash (256mbit when AMD was at 64mbit) to trailing.
Every mirrorbit sale has came out of the hide of Intel's formerly lucrative high density flash ex-monopoly.
Re: Looks like you're wrong, Dan. Again.
Better look at that list, again. Most of the flash FABs are mixed logic and flash.
Re: I asked for a pointer that says Intel is going down the tubes.
Sitting on a pile of cash but with steadily shrinking market share?
Having to borrow money for the first time in many years?
You said they were going down the tubes, those weren't my words.
But I certainly won't argue with you...
Re: Fab 23 = Flash only, fab 17 = logic only
Our favorite link!
But it also lists 11, 18, 8, 10/14, and D2 as being mixed logic (chipsets, com parts) and flash.
Besides, Elmer used to tell us over and over again that he worked in the FABs and that Intel's FABs all ran mixed product types - surely you wouldn't call old Elmer a speaker of falsehoods?
Re: there aren´t even any numbers that show that AMD gained share from Strataflash with Mirrorbit
Keith, almost all mirrorbit sales represent conquest sales from strataflash.
AMD did not even have parts in the 128 and 256mb segments until they began shipping mirrorbit - it was a 100% Intel market.
Until AMD began shipping mirrorbit, high density NOR equaled Intel strataflash.
Re: Sure, growth from 0% to anything carries something like infinity percent jerk. Of course, StrataFlash is also growing, albeit at a more mature pace given that it's a mature technology
Did a new guy get assigned to the wbmw desk in the Intel PR department? You're both less informed and less articulate than the last guy.
MirrorBit™ Technology Timeline
2001: Announced revolutionary new Flash memory technology
2002: Ramped first product to mass production
2003: Ramped first full family of products (16 to 256 Mb) to mass production
2003: Received INSIGHT Award for Most Innovative Memory Device
2003: Introduced world’s first 512 Mb NOR Flash memory on second-generation
technology
2004: Introduced industry’s best price-performance Flash memory for wireless
using second-generation technology
2004: Ramping second generation technology to mass production
2005: Introduction and production ramp of 90 nm MirrorBit™ technology
including solutions for high-density media storage
http://www.spansion.com/technology/2004_MirrorBit_FUJ.pdf
Re: I have a few posts bookmarked so I can throw them in their face when reality sets in and Yonah turns out to be a success
Be sure to include the ones about flash and chipsets coming from the same FABs.
Really rub it in.
wbmw wrote: This one is your imagination. Chipset and flash production is totally separate. They are on two different processes and use completely different fabs.
You should get a clue.
SANTA CLARA, Calif., Sept. 15, 2005 -- Intel Corporation today announced plans to invest $345 million in two of the company's existing manufacturing sites in Colorado and Massachusetts. The investments will be used to increase the capacity of the wafer fabrications facilities (fabs), Fab 17 in Hudson, Mass. and Fab 23 in Colorado Springs, Colo.
Both fabs are 200mm facilities that produce primarily chipsets communications and flash memory components for a variety of Intel platforms.
http://www.intel.com/pressroom/archive/releases/20050915corp.htm
Re: "I believe it, so it is true"?
Intel produces CPUs at 5+ FABs but most current product is from D1C, F11X, and Fab 24. Call it 3.5 FABs. 12" FABs support 50% to 100% higher wafer area throughput (wider wafers, but they take more time to process) than 8" FABs.
Intel's clean rooms are 150 to 200KSF averaging 175 vs. AMD's 150KSF.
So Intel has 3.5 (no. of FABs) * 1.5 (adjustment for 8" vs. 12" wafers) * 175/150 (larger clean rooms) = about 6 times AMD's CPU capacity, yet AMD is gaining share on Intel, and the ratio in the market is between 4 to 1 and 5 to 1 (and AMD looks like they're selling out and need the new capacity they've just brought on line).
Judging by share, and some Intel claims of full, or nealy full utilization, AMD has better yields than Intel. If Intel is even close to using all of its CPU capacity, then that still holds true. Average Die sizes would impact the calculations above so please feel free to take a shot at it. I think the average die size is pretty close, because the bulk of Intel sales is from a mix of ~90mm2 (Pentium M) and 115 (Celeron/Pentium) While AMD has a mix of Sempron (~85) and Athlon (~115). AMD's dual core parts are ~145 while Intel's are ~200 but my impression is that AMD's dual core sales are a larger share of their total, such that overall average die size pretty close for the two companies.
AMD is doubling+ the number of FABs it has with the output increase first showing up next quarter (too bad for them it wasn't a quarter earlier). Intel is adding a 4th main FAB.
AMD capacity is increasing 100% next year, while Intel is increasing 33%, but dual core parts and larger caches are acting to increase die sizes for both companies (despite the move to 65nm) so the total increase in joint capacity is smaller.