3.27.2007

Intel May Win Its War With AMD: Long Term Road Map Looks Promising

From Seeking Alpha
http://chip.seekingalpha.com/article/30791

TWP notes today's CPU environment is a difficult one, particularly with the aggressive ongoing price cuts between Intel (INTC) and AMD (AMD). The firm is incrementally more positive on Intel, however, given their findings regarding its long-term server road map, and particularly the Thurley platform. They believe that the dramatic change in architecture (i.e., integrated memory controller and CSI) bodes well for the company as it transitions off of the Bensley platform in 2H08...

Intel will introduce its next-generation server platform named Thurley in 2H08 and will feature the Gainestown processor and Tylersburg chipset...

Gainestown-Intel's masterpiece in the works: Similar to Barcelona, Gainestown will be a monolithic quad-core processor in the Nehalem family and will be manufactured at a mature 45nm node.

Once again, Scientia is incorrect to say that Nehalem will be introduced 2009. He once questioned my logic that Nehalem will be released just be in time for the Barcelona launch. What he keeps to be in denial is the fact that AMD already confirmed no significant Barcelona ramp until 2008 which really coincides with AMD's 45nm transition. AMD promised to start its transition on 2H'08. And guess what Intel is launching at the same time? Anybody who's done work in a Fab, especially in a constraint one, already knows that a monolithic QC at 65nm isn't cost competetive and would definitely be a barrier in trying to gain market share.

As far as AMD's execution is concerned, we're assuming that AMD keeps its promise of a 2H 08 45nm transition. Judging by experience, AMD's 65nm transition which was promised 2H'06 actually happened just a few days shy of 2007. And of course we're also assuming Intel doesn't pull in Nehalem which it recently did with Penryn.

All in all, AMD will ramp a new micro-architechure just as Intel releases a new one. Barcelona will never have the opportunity to shine like it did with the Opteron. K10 will be an old architecture even before it gets the chance to ramp.

3.26.2007

64bit NUMA benchmarks confirms AMD's obsolescence

Finally a comprehensive benchmark that proves that AMD's Athlon doesn't work better in 64-bit mode or with the NUMA enabled Windows Vista. This is yet another painful blow for the aging microarchitechture from AMD.

From the Tech Report:
CONLCUSION: The fact that Intel retains the overall performance crown comes as no surprise. As we said at the outset, AMD has no real answer to the Core 2 Extreme X6800 among its dual-core processors. Also, Intel's quad-core CPUs tend to scale better than AMD's Quad FX platform, especially for typical desktop-class applications. Our move to Windows Vista x64 has done little to alter this dynamic. At the same time, Core 2 processors tend to draw less power and to be more energy efficient—sometimes markedly so—than Athlon 64s. Right now, Intel has the magic combination of a superior processor microarchitecture and a more mature, fully realized 65nm manufacturing capability working together on its side.

3.21.2007

AMDZONE against the Coalition of the anti-AMD

One must feel sorry for AMDZONE News poster Chris_Tom. Lately his role has become like Mohammed Saeed al-Sahaf, the Iraqi information minister prior the invasion of Iraq. For Chris, the reality of the situation about AMD’s decline, both in terms of performance and finance doesn’t matter. For him all that matters is what he says and gets to say it last.
The barrage of negative news about AMD has turned him against just about everybody who has nothing good to say about AMD. The same analysts whom he loved to quote a year ago when it was all about AMD’s rise can no longer be trusted today. The same review sites which he posted 2 years ago showing AMD’s K8 taking the lead are today either biased or under Intel’s payroll. Today he’s after SGI.

SGI has pushed a press release about a new Xeon setup they sold to International. Frankly I'm surprised SGI is still in business, but putting flat out lies as a press release won't help them. So now Opteron is 3-5 times slower than Xeon eh? The press release fails to mention any details about the Opteron setup they are replacing. A 3-5 times faster mention makes this an Apple press release. Are Apple spinners doing the same for SGI? It sure sounds like it. The facts are head to head tests of Opteron and Xeon, in the latest incarnations clearly show that there is no 3-5 times gap in performance for either CPU. That is just nonsense. As a 77 International Scout II owner I'm pretty disgusted that International has been dragged into this one. “

And just like every single one of his evangelical-style rebuttal of anything bad about AMD, he doesn’t offer any argument. All he does is attack the news writter calling them a liar, incompetent or clueless. Why should he? His readers don’t seem to demand counter arguments. Maybe attacking the messenger and not acknowledging the message gives its readership the comfort and the illusion they need in this crisis. Do they honestly think this will alter the reality of AMD’s downward spiral? I have no idea what they think. But I do know what happened next when Saeed al-Sahaf repeatedly said there were no American soldiers in Iraq.

3.19.2007

AMD Seems Alone in its Price War

“Banc of America affirms caution on AMD, and cuts their tgt to $14 from $15 and estimates. The firm's channel checks indicate that AMD is preparing for another round of price cuts on Apr 9th, in an effort to 1) bolster flagging demand for its products and 2) preempt Intel's price cuts scheduled for Apr 22nd. These price cuts, which they believe will be across of several AMD's desktop products (including Semprons, Athlon X2s), and will range between 26-42%. Regarding Intel (INTC), while they think the price cuts will put some pressure on Intel's low-end offerings, they expect blended pricing for Intel's desktop processors to trend flat with higher volumes in the premium Core 2 Duo segment offsetting price erosion in lower end processors; the stability in notebook and server pricing (no price cuts) should serve as an additional mitigating factor.”

Once again, market researchers confirm that AMD is in a price war with Intel’s low-end products only. Meanwhile Intel continues to milk the high-end cow completely unchallenged offsetting margin erosion from its price-skirmish with AMD. The Centrino brand continues to sell while Xeon continues to take back market share from AMD. Soon AMD’s OLPC plan can come into fruition when then have to give away their X2’s for free just to clear inventory. With Barcelona’s limited volume we can expect AMD to bleed red until they can volume produce at 45nm – 2009 to be more precise.

3.15.2007

AMD's Chipset Business is Insignificant

From this article:

"We don't think the 690 chipset is a bad chipset, it is something that is good but not something that would change the game for the company," said Wang. He added that his firm is forecasting that the AMD 690G chipset is expected to make up three to four per cent of AMD's sales during the first quarter but is seen as a minor catalyst against larger issues such as competition from similar chipsets from Intel and NVIDIA.

Once again, this shows how completetly OFF some of the pro-AMD bloggers are. Let me repeat what I said. Chipsets are "enabler-parts" and businesses like AMD and Intel with gross margins hovering around 50% do not make these parts for the sake of profits. You make them to create the environment where your bread and butter can exist.

Anyone expecting AMD's chipset division to save the the company from its current predicament clearly doesn't know what he's talking about.

3.09.2007

“Everything that can go wrong has gone wrong” for AMD

From the Houston Chronicle:
“Chip maker Advanced Micro Devices Inc. may be losing market share, while continuing to face pricing pressure and cash flow challenges, according to Friedman Billings Ramsey analyst Chris Caso.”
"Our checks and AMD's own comments suggest that the company's first-quarter miss is related to share loss, not market softness," Caso wrote in a note to investors. "In contrast to AMD's comments, however, we believe this is more than a one-quarter problem, due to excess AMD inventory at customers, what we expect will be continuous competitive pressure form Intel, and AMD's urgent need to raise cash."
Caso reiterated an "Underperform" rating on the stock, and said he thinks the stock could drop to as low as $10.


AMD’s uncompromising marketshare strategy while sacrificing everything that matters in the business world like turning a profit seemed to have grinded to a halt and failed miserably. Now AMD has nothing left but a very big grave that they seem to have efficiently dug over night. Missed earnings, alarmingly low margins, poor product offerings, slow execution, huge debt, customer inventory, expensive and inefficient Fabs and poor yields. The future doesn’t look bright either. Any chance of AMD at least increasing ASPs with the release of K10s is now completely gone with the aggressive pull in of Intel’s Penryn 45nm family.


When AMD described Intel’s admission of losing e-mails as “everything that could go wrong did go wrong”, it seems that this sort of phrase is commonly used internally by everybody in AMD when reporting updates.


I sure hope that Scientia will now admit how he severely underestimated Core2Duo and he was completely mistaken to think that 2007 will be a good year for AMD. I honestly don’t know what else can AMD screw up to change his outlook. But I'm not holding my breath.

3.05.2007

AMD to miss Q1 Targets

From Businessweek:
"Chip maker Advanced Micro Devices Inc. warned Monday it is "unlikely to meet" its revenue guidance for the first quarter.
The company's outlook was for sales of $1.6 billion to $1.7 billion.
Analysts polled by Thomson Financial are looking for $1.7 billion in sales.
Shares of the Sunnyvale, Calif., company were down 48 cents at $13.69 in premarket trading, after closing Friday at $14.18 on the New York Stock Exchange."


Bear in mind that these numbers were already guided down from the last, already disappointing, conference calls. This is not a straightforward comparison from the previous year due to the inlcusion of ATI's revenue but it should be common knowledge that both AMD and ATI are lossing money with its luckluster product offering and poor execution.
The question now is, by how much and what impact will it have on any propects of raising the badly needed money.

3.01.2007

AMD cries unfair while Intel takes server share

From the Register:
"Intel and AMD have embraced a divide and propaganda approach to tout their upcoming server chips. The two vendors today held press events in San Francisco at the exact same time, forcing journalists to pick between the giant and the underdog... Top reporters from the likes of the Wall Street Journal and CNET made the same decision (ignore AMD). Not AMD's finest hour from a planning standpoint, as you can tell."

It's hard for journalists and enthusiast to be excited about anyting AMD has to say. Intel has been coming out with impressive demos and aggressive products. Conrast to AMD who been announcing losses, delays and releasing sub-standard parts. Even its up and coming Barcelona hasn't been seem running anything. It's no wonder nobody is interested.


"The company's collection of some 20 Opteron-beating benchmarks prove that AMD will struggle against Xeon for months to come – even when it releases the four-core Barcelona chip – according to Intel's VP Tom Kilroy.
AMD has charged that Intel's benchmarks are unfair because they compare the latest and greatest Xeon against an older Opteron. To that claim, Intel says, "Pah." It always compares against AMD's freshest published results. AMD ought to publish new benchmarks more quickly if it wants to play.
Without question, the Xeon line looks pretty impressive these days, despite Intel's reliance on aging technology such as the front side bus. Intel claims that it now takes 9 out of 10 new server wins on Wall Street – quite the change from last year."

Understandably AMD doesn't want to publish their latest results because that would take away their excuse that Intel is unfair in comparing its latest from AMD's older Opteron. Not that it really matters because all major review sites have already demonstrated how the Opteron lags the new Xeons in real life applications - by a very substantialy margin. AMD's whining will not help stop Intel's rate of 9/10 new server design wins. They need to release Barcelona before Penryn comes out otherwise they would look like they're not even trying to catch up.