Intel to delay Yorkfield chips because of AMD's struggles?

Chalk this one up to wild unsubstantiated rumor, but Digitimes is reporting that Intel may hold off on launching its 45nm quad-core Yorkfield chips -- which were supposed to hit on January 20th -- because it's not facing any threat from AMD's delayed Phenom chips, and launching now would just eat into existing 65nm chip sales. That means we wouldn't see the Core 2 Quad Q9300, Q9450, or Q9550 until Intel feels the heat from AMD, which seems like a pretty timid strategy for a company whose former CEO's motto was "Only the paranoid survive." Add in the fact that a certain fruit company is rumored to be launching new Mac Pros with some version of 45nm Penryn chips next month, and we're really not feeling this one, but only time will tell.

[Via The Inquirer]

[why not the LS2LS7? @ Dec 19th 2007 1:11PM]

My understanding is Intel has having some yield problems at 45nm right now. It's from a semi-reliable source.

If they intro new chips at the top end and the market tries to shift to them and they can't deliver them, it's a problem. Better to just leave the situation alone until they can intro a new chip and make it available in proper numbers.

[Totalfixation @ Dec 19th 2007 1:23PM]

That was what i was thinking, make the competitor look bad, even though they are having problems of there own.

[Louis Duran @ Dec 19th 2007 4:18PM]

I would delay if I were Intel especially if the yield numbers are not quite as good as the 65nm product. Intel reads the news just like all of us and they saw what happened to AMD when AMD launched a faulty processor. Intel are smart enough to know that if they had as bad a launch as Phenom, it would be much worse since Intel are EXPECTED to have a solid launch and AMD is EXPECTED to screw up a launch. It is all about expectations from the customer. Intel just have higher standards than AMD.

[WeirdCrankP @ Dec 19th 2007 7:16PM]

I think it's not so much yield issues - those are mostly taken care of. Yield was bad back in November, but has come up significantly since then.

Reason to wait: Intel has a chance to sell its backlog of 65nm chips. Also, the new Arizona fab isn't producing yet and the Israel fab is far behind schedule (= won't produce until the end of 2008 at best). Right now, only the Portland fab is producing 45nm chips. Waiting would give everybody some breathing room.

That, however, doesn't sound like Intel at all. If they can squash AMD, they will. That's why I think the report is BS and the chips will be out in Janurary as planned.

[Matt B @ Dec 19th 2007 1:20PM]

I'm sure that Intel holds some stock in AMD and don't want that stock to be worthless anytime soon. I was always a big AMD supporter in the past. But I know my next upgrade will be from a 3800 X2 to an Intel dual core.

[kempcross @ Dec 19th 2007 2:52PM]

Why would Intel own any AMD stock?

[Mr.Tech @ Dec 19th 2007 5:45PM]

He's right... seriously why would Intel buy any of AMD stocks? Wouldn't Intel rather buy back it's own stock =D

[Matt @ Dec 19th 2007 8:37PM]

Why would Microsoft have purchased Apple stock?

[mypetridish @ Dec 20th 2007 5:03AM]

for the same reason Yahoo bought Google stock.

it was a good decission. companies buy competitions' stocks all the time.

[Ray-- @ Dec 19th 2007 1:30PM]

I thought moving to a smaller manufacturing process saves the company lots of money (i.e. more chips per wafer). Why on earth would intel stay at 65nm if it costs them more? The more plausable conspiracy theory (if we are inventing them) would be that they would move to 45nm but slow the chips down some... so that later on when AMD came around they could simply boost them up to full speed to squash their release.

[Ed @ Dec 19th 2007 2:41PM]

If they are having yield issues, then maybe they can't make as many functional chips on a 45nm wafer yet. 65nm wafers probably have great yields by now.

[Poom @ Dec 20th 2007 12:54AM]

Perhaps Intel already produced bunch of 65nm chips and want to sell them out first before they start selling 45nm, which will reduce sales of 65nm chips.

[Alfred @ Dec 19th 2007 1:42PM]

that doesn't make sense, lets wait for our competitor to threaten us before we move forward, forget that if you have the ability to completely knock your only fight out of the water do it. Its like saying that in a boxing match the obvious winner is going to take it easy and wait for the other guy to throw some really solid blows before he starts using his energy

[pkilla @ Dec 19th 2007 2:14PM]

Actually it makes a lot of sense. They dont want to completely stop AMD, if they do the they can be brought up on monopoly talks. Businesses needs competition even if it is no contest who is better.

[indoorsy @ Dec 19th 2007 2:56PM]

Didn't you ever see Rocky III?

[AlfredTV @ Dec 19th 2007 5:19PM]

that's true, didn't think about the monopoly issue. but at the same time the could get amd to a point that they could buy up the maximum amount of stock to have a key stake with out having those issues and then help build them up again so then they would pretty much own amd and do what they want....

[Matt @ Dec 19th 2007 3:27PM]

If there's no competition from AMD, they're just going to cannibalize their 65nm sales, like the article says. It's best to wait till you can eat someone else's lunch.

[Chuckles McGee @ Dec 19th 2007 6:47PM]

If Intel's much better than AMD, there's no point in introducing higher cost chips at the same pricepoint. People who want a new chip will go with Intel's fastest anyway.

[animeGhost @ Dec 19th 2007 1:43PM]

well... i heard a while back that intel and amd have some sort of development contract were one can backwards engineer the others chip and use some of their technologies (not sure how true that is, i'll have to research it). My main theory is that intel cant handle amd going down. not only would it make them a monopoly prone to all sorts of anti-trust laws but there's no way they'd be able to keep up with the demand for basically all the processors that come in new PCs and in Macs. (just my theory)

[Silverfrog @ Dec 19th 2007 3:11PM]

AMD isn't going to die anytime soon. They've had troubles before, when initially challenging Intel back in the 90s. This is around the time when the K6-2 and K6-3 chips were out and doing rather poorly against the competing Pentium 2 and Pentium 3s. That is--until the Athlon was released, then the game changed. And, the rest is history.

This is the first time in a long time when AMD wasn't on top of their game, but realize they will rise again. In my opinion, their architectural strategies are superior to Intel's. Intel just shrunk their die size down and is just now about to integrate the memory controller on-die. Meanwhile, AMD is moving forward with TRUE quadcore (4 cores on one die as opposed to Intel's two dual cores on one die design). Right now, Phenom doesn't have the clockspeed or horsepower to compete in the high end. This is also hampered by the lack of DDR3 memory.

But, this is most likely the calm before the storm that AMD will undoubtedly unleash in the next 6-12 months. We will soon see graphics cores on-die with CPU cores, which will means possibly one day eliminating graphics cards altogether in favor of unified interfaces.

Even [if] AMD did go under, IBM is quite strong and could very well take up the remaining market share that AMD left open for grabs. Let's not forget that the future is going to be a star-trek based light/optical transmitter-based CPU, using photons to carry information instead of wires. Less power required, they're faster, and generate less heat. IBM could be a player, come 2015 or so. Wait and see.

[Louis Duran @ Dec 19th 2007 4:27PM]

@Silverfrog
"Let's not forget that the future is going to be a star-trek based light/optical transmitter-based CPU"

OK Dude... step away from the bong. I heard that the licensing of that star-trek based CPU technology hasn't been going so well ever since Gene Roddenberry died. :-)

Seriously, I don't think in 7 years microchip process technology will change significantly. Maybe we'll be at 15nm litho-etch still but light photons being competitive by then? No way. They may be able to demonstrate the technology but mass produce... and compete on performance with semicondictors? No.

[Silverfrog @ Dec 19th 2007 4:40PM]

Re: Louis

All I meant by mentioning Star Trek is that it had the idea years ago with the "neural-net" computers. I'm sure Trekkies can attest that many inventions are based loosely off of the series.

Other inventions, namely the transporter among other things, has also been conceived and theorized. read here.

http://domino.watson.ibm.com/comm/wwwr_thinkresearch.nsf/pages/explorations198.html

Again, my post only insinuated the mere possibilities of using light/optical based CPUs. it's only a prediction.

[Silverfrog @ Dec 19th 2007 4:47PM]

Re: Louis

Also, here's the link to IBM's research for Optical storage and computing, in case you still don't believe it's possible.

http://www-128.ibm.com/developerworks/blogs/page/powerarchitecture?entry=new_buffer_uses_photons_for

[Brian @ Dec 19th 2007 2:20PM]

I guess there is something to be said for competition after all.

[Teetdogs @ Dec 19th 2007 2:48PM]

Well the whole monopoly thing wouldn't happen I dont think since there are other manufacturer who make microprocessors even if they aren't used for run of the mill desktop computers.

[eesz @ Dec 19th 2007 3:47PM]

If it were not for AMD, Intel would still be selling us Math Co-processors.

[tekdemon @ Dec 21st 2007 3:26PM]

I don't know why some people seem to think that AMD somehow is the sole reason for Intel ever bothering to innovate.

There was a time when AMD's products were not at all a threat to Intel, and Intel always innovated anyways just to get people to upgrade from an older Intel chip to a newer Intel chip.

Without AMD, Intel would still have to convince people to buy a new processor every couple of years. Which is why they introduced Pentium MMX to succeed the original Pentium, why they designed the P6 architecture, etc. During that entire stretch of time all AMD had was either a 486dx100 or a K5 (with really awful yields on the K5). It wasn't until K6 that AMD was realistically a competitor.

And anyways, once AMD became a serious competitor with K6-2 and Athlon, Intel responded to this competition with Netburst, so maybe Intel is better at innovating when they have to actually improve upon their own chips instead of getting into idiotic Ghz marketing wars with AMD.

[Daven @ Dec 19th 2007 3:54PM]

I'm sure the writer of this blog post knows that all Mac Pros use Xeon processors and not the desktop variant. Xeon versions of Penryn were released 2 months or so ago. You can buy Quad core 45 nm Xeons from 2.0Ghz up to 3.2GHz in FSB 1333 MHz and 1600 MHz versions. Therefore, said rumor will in no way effect Apple's Mac Pros AT ALL. It is rumored that Apple is just waiting to stockpile enough of ALREADY EXISTING Penryn Xeons before releasing new Mac Pros.

I wonder why these tech blogs can't differentiate between Xeons and desktop variants or get the fact that NOT ONE single Apple computer uses desktop variants. They only use the mobile Core 2 (iMac, Macbook, Macbook Pro, Mac Mini) and the server/workstation Xeon (Mac Pro).

This was a problem when Apple was about to first release Intel Mac Pros. Everyone was guessing which Conroe (Core 2 desktop variant) Apple was going to use in the Mac Pro when everyone already knows that Apple uses dual processor configs for the Mac Pro. That means only Xeons will work.

Alas maybe one day these tech blogs will get it.

[Shadowise @ Dec 19th 2007 4:39PM]

You can't really blame Intel for doing it. Obviously, this doesn't benefit the consumer, but I can see why they wouldn't release something new, when their current lines are already walking all over the competition. AMD really dropped the ball.

[Redsyrup @ Dec 19th 2007 5:00PM]

Let's not forget that competition drives innovation, without AMD, Intel becomes a lumbering giant. This could just be the begining.

[Ian @ Dec 19th 2007 5:23PM]

Sounds like a prudent corporate decision. Get the 45 nm yields up nicely, by allowing a further 30-60 days and meantime sell more high profit 65 nm chips. And it doesn't push down prices of those 65 nm cpus either. A win, win for Intel. We the consumer lose out slightly because AMD have fallen so far back. In a back handed way it actually provides some relief for AMD, with Intel declining to go in for the final "kill". All those attacking Intel's "monopoly tactics" will now also be forced to eat those words, by Intel backing off like this. I am sure Intel knows this as well. AT the end of the day Intel needs a competitor so slaying AMD by selling so much better chipsets isn't in their game plan. IMHO. The EU especially should take note and say "thanks".

[pilot101 @ Dec 19th 2007 5:40PM]

damn amd, now is the chance if the opposition is providing you would chance fire all the problematic managers, shape up and kick some serious butt again ^^, maybe launch the gpu and cpu connected chips a year before schedule, like next year, instead of 2009.

[LeroyLeroy @ Dec 19th 2007 7:25PM]

I thought this delay was reported a couple of days ago, as a favor to motherboard manufacturers:

"Intel may postpone the release of X48 by 1 or 2 months -- Intel had been planning to launch the new chipset in January -- because one major motherboard maker still holds a large volume of X38 chipsets in inventory, according to a report in DigiTimes."

http://money.cnn.com/news/newsfeeds/articles/newstex/AFX-0013-21720533.htm

[AdamRLeggett @ Dec 19th 2007 8:24PM]

Assuming there's some truth to this - is it possible Intel just want to delay the official launch of Yorkfield so it can provide Apple with all the chips it wants - all without having to say its available in limited quantities to everyone else.

[Alejandro Luciano @ Dec 19th 2007 8:33PM]

cell phone

[yannick @ Dec 20th 2007 12:10PM]

here we go again, intel is lazing about and isnt isnt gonna release new procs, it is the p4 age all over again. itel r&d is asleep again, until amd wake them up.

[palehorse @ Dec 20th 2007 3:00PM]

It would be just like those arseholes to delay MY rebuild! Damn them! Damn them all!

:)

[peter Stadler @ Dec 20th 2007 4:08PM]

It is an easy decision for Intel. They have huge resource for 65nm production and they habe limited resources for 45nm production and they have a lot of new chips design for 45nm production only.
Intel may produce 10 Silverthornechips instead of one quadcore chip and because of the lack of competition there is no need for hurry up the transition from 65nm to 45nm for certain chip lines.

So the real question is, which chips will Intel produce instead of the delayed quadcore chips