Too early to read AMD last rites: analysts
Analysis Won't go gently into that good night
AFTER RACKING UP losses of almost $4 billion last year, shedding thousands of staff, including several senior execs, and with its stock price seemingly in a tailspin, you could be forgiven for wondering whether AMD actually has a future ahead of it.
With intensifying competition from rival chipmaker Intel putting the firm under pressure, and feeling the pinch from the $5.6 billion it coughed up to buy graphics chip maker ATI, AMD has definitely been feeling the heat of late.
Many looked to the company’s annual shareholder meeting last week to shed some light on just how AMD hoped to dig itself out of what appears to be an ever-deepening hole. But chief exec Hector Ruiz was saying nowt.
Would AMD be going “asset light”, or just “asset smart”? Was there a master plan in place to pull the chipmaker out of the doldrums? Ruiz wasn’t going to say. At least by keeping quiet, no further damage could be done, and it didn’t hurt the stock price to inject a bit of mystery and intrigue into the proceedings either.
No-one denies that AMD has had something of a year from hell, but talk of the chipmaker’s imminent demise don’t cut much ice with analysts.
Although the delays in launching Barcelona did much to hurt the firm’s reputation and credibility, few believe it was the final nail in AMD’s coffin.
Chip-spert Nathan Brookwood of Insight64 told the INQ that although he was “disgusted with the way AMD ran Barcelona over the last 12 months”, he didn’t b elieve it would spell doom for the company, especially now that Barcelona was actually shipping, and firms like HP, Dell and Sun had all launched products incorporating the chip.
Rob Enderle, principle analyst at Enderle Group, said Intel had also missed past launch dates, but it had recovered well and AMD probably would too. Even so, AMD had “better not do it again anytime soon or they may not recover”.
“AMD basically have no credibility now and the only way to get that back is to deliver,” said Brookwood. AMD wasn’t exactly languishing in the pit of despair, and the chipmaker was “pretty confident it can deliver”, he added. Getting its Shanghai 45nm server chip out on time would definitely be a step in the right direction.
Fabless or fab-lite?
As to whether AMD would go fabless, opinions were divided.
Jack Gold, head analyst at J.Gold Associates doesn’t believe AMD will sell off its fabs, but was likely to rely more heavily on outsiders to do its manufacturing. AMD already allows TSMC to manufacture a significant number of its chips, but because TSMC’s fabs are at least one-and-a-half generations behind Intel, offing the whole manufacturing process to TSMC seemed an unlikely move, he said.
Gold noted that AMD had to “be a lot smarter in how it optimises products” being at the mercy of a manufacturing facility the company did not control. “This is no easy task,” he said.
Brookwood was also sceptical about AMD going taking the fabless route, saying he thought it was more likely the chipmaker would be “looking for partners” such as Singapore’s Chartered, TSMC or possibly even IBM.
When it came to running fabs and fab technology itself, Brookwood said AMD “came top of the class” which would make the company extremely reluctant to get out of fabbing.
“Besides, there is not really much point in selling a fab if you have enough business to keep it going, and fab 36 is running at close to full capacity.”
Enderle, however, sees things differently. Giving up its fabs would give AMD “flexibility” and would “reduce significantly its capital commitment, while making better use of in-place resources”. In the short term, though, he expected the chipmaker to be relatively careful with how much fab capacity it gave up, learning hard lessons from Transmeta and Cyrix.
There has been much speculation that AMD might have to sell off its Dresden fab in Germany in order to raise a bit of cash, but the issue is not exactly straightforward as AMD was given loans by the state of Saxony to build the fab in the first place.
As Broockwood points out, AMD did still generate cash last quarter, and despite its losses, was still turning enough profit to avoid panicking and frantically selling off everything it owns.
It is no easy task to simply shut down a facility, noted Gold, saying it would be especially difficult to do so in Europe unless AMD sold it outright to a new owner who would take over its operation.
“With the current state of the semi market in Europe, I am not sure anyone would want to purchase the facility outright and continue to operate it”, he said.
Shake, rattle and heads a-roll
Most observers agree that the ongoing management shake-up, which has seen the
exodus of at least four senior execs – most recently Mario Rivas, executive VP
of computing solutions and chief talent officer Michel Cadieux – were part of a
drive to get the compay back on track rather than a dash to abandon the sinking
ship.
“Sure, AMD have lost a few people, but they’ve also been reorganising”, says Brookwood. “I get the feeling that some may be being pushed overboard rather than jumping ship.”
Gold agrees that not all the departures appeared voluntary.
Both of them seem to be referring to Rivas, who might well have been forced to walk the plank over the whole Barcelona launch debacle.
Enderle said the changes were “a good thing” as the company sorely needed new blood to shake things up and get it back on track.
A graphic future?
On the subject of AMD’s purchase of graphics firm ATI, both Brookwood and
Enderle said the acquisition, although expensive, was an important move, and one
that could set AMD up as a significant Intel rival by positioning company
efforts against Intel’s Achilles heel -- graphics.
“It is a costly move, but in five years we may look back and realise it was what AMD had to do to gain some kind of competitive advantage” explained Enderle. “It was very risky, though”.
Brookwood said ATI was at a very good point in its product cycle, and its acquisition was wise because it gave AMD “access to a great big pot of intellectual property”. Did the company pay too much for ATI though? Brockwood doesn’t think so. “They negotiated and came up with a price that was acceptable to both”, he said.
“AMD can not compete with Intel across the board on all products”, Gold said, noting that AMD had to “be smart and innovate where it believes a niche exists that Intel does not fill”. This is why a move into the graphics and visualisation areas could be so critical for the company.
No one denies that AMD has had something of an annus horribilis, and that there is still much work to be done to get the company back on track. However, it would seem there is light at the end of the tunnel, even if the company doesn’t go fabless.
AMD says it will ramp up production in the first half of this year, which probably means chips should be ready to ship to OEMs by September, and OEM products due out in either October or November.
AMD also has to start being more cautious with its investments seeing as the company, though far from going broke, doesn’t have masses of money to spend at the moment.
Although AMD is unlikely to recapture 25 per cent market share when it comes to servers, it may still be able to recover a fair bit, because at least now the products are available.
“The company is not sending out unarmed soldiers to the battlefield to do battle with Intel and its canon,” Brookwood says.
“AMD has to be very cautious, meet its commitments, do whatever it can on the core, on the chip, and increase performance.”
Then it will be doing just fine. µ
See Also
AMD to outsource more CPU production to TSMC - report
AMD's Dirk tells the troops what is happening
AMD's asset-light hype boosts Chartered Semi
AMD's "asset-light" plans hotly anticipated
Comments
ASBOtastic Genius!
DAAMIT a grizzly bear!$5.6 billion doesn't nearly buy half what it used too!
Down boy! Nice teddy...
I think we should be on to playing dead!
Analysts HAVE TO keep the dream alive
Whatever.(Posted in reply to another on Inq article comment today):
AMD is sucking hind tit in both CPU and GPU, that's fact, and with all the problems on their plate, AMD don't have a reasonable out.
Playing Intel's game by Intel's rules has kept them in 2nd place since their first 386 clone. No one at AMD has ever understood they can't win this game, they just all dream of leadership glory without ever really knowing how to get there. It takes leadership and vision people. Now that Intel is back in top form that game isn't going to change any time soon, AMD will remain at 15-18% CPU market share because Intel allows it.
On the graphics front, it's game over. AMD's purchase of ATI has assured there is no opportunity for revolutionary or even evolutionary performance gains to beat NVIDIA; AMD has too many other problems to allow graphics to be their leading technology, and NVIDIA just continues to execute. Note the exodus of graphics tallent from AMD right now. Who is going to build great products for them? No one, they've all gone.
Is AMD going to survive? Sure, it wouldn't be in anyone's (Intel, major OEMs, the rest of the industry's) interest to let them fail.
Are they going to thrive? No possible way, if you're not growing you're dying. Why?
a) Hector is an idiot and couldn't manage his way out of a wet paper sack, and killing off all his underlings isn't the problem, try using a mirror Hector.
b) Debt -- something tech doesn't do -- is AMD's balance sheet highlight and is akin to an anchor tied to a struggling swimmer, unless a big sugar daddy steps up, debt service is going to hold them down for a long time.
c) Leadership and product innovation is gone in CPU with no way to regain it
d) Gaining an edge in process technology is too expensive at this point due to the debt picture.
e) Personnel - tallent is leaving in droves, leadership/visionary folks like Henri Richard are gone, moral is at all time lows, and besides, AMD has about 25% too many staff based on their cost structure.
f) The last great hope, Fusion, is designed as an entry, commodity solution -- no one (including Intel) is going to get rich building this type of product, and it certainly isn't the answer to all AMD's woes as it has been positioned.
Besides that, the game is changing -- x86 has run it's course. The margin-rich segments like HPC are going to be vastly different in the next 5 years than they are today. And the standard x86 notebook/desktop experience we all have come to know and love is going away. Desktop computing is going niche into areas like gaming and workstation, and business/home computing is moving to something you can put in your pocket. These areas are where AMD needs to be innovating and they certainly have the assets to make something happen there. But you know what? THEY CAN'T AFFORD TO INNOVATE HERE. They don't have the cash or cash flow to fund the billions of dollars of investment it's going to take to be a leader.
Now, all the near-sighted fan bois can rah rah all they want, but the smart ones see the writing is on the wall. Hell, I've owned a bunch of AMD systems over the years, and AMD will not go out of business in the near term, but AMD are now and will remain a shadow of their former selves with between little and no hope of recovering past glory.
Hector and the debt have to go before anything meaningful can happen.
Mirror,Mirror-Anos Prophetics.
AMD Might be in last Few Public rounds before going Private in Ow, Sai 2018. Or AMD Could use Underdog Trick, Grab DRM by Reproductive Thangies & quash it.If design gets too complex & their all Flexing for that info, Back to Time & Charges plus about trillion other things.
Or All-In-Wonder that records BluRay easy. First put it in Cpu+Gpu combo so have to BUY New Computer, Hahahah. Then when Amazed Public is at: HOW Brazen Prirates Can live, Go to All-In_Wonder crowd, as cpu stuff is so complex its AMD Exclusive.whyUK?loc
Then finally hit world with plenty of basicly discontinued All-In-Wonder cards out as Public has Incredible Viewpoint there worth more TOP End Cards. Wow, Peepers.
Finally BUY Me Castle in US/UK Zone. & make em all sprenken der Englishe'.
Thomas Stewart von Drashek M.D.
Blah
Before Barcelona was out AMD was quiet as a fish. Analyst were being as quiet.Now all they do is talk and talk talk.
Your last chip was a year late & under-lcokced.
Enough talk, no apologies, no analysis. Bring forth the chips, let us see if the second pitch is a strike or a ball.
This Play has only one ending
The lady doth protest too much, methinks.AMD in trouble
All this is assuming that everything comes up roses for AMD.What is the guarantee that their new projects will do any better than Barcelona is doing?
They were confident there as well, and we can see what happened. They have no new IP for this year at all. Everything they do will still be a slightly adjusted K10, which, after all the hype, is nothing more than a warmed over K8.
I'm skeptical.
Selling Tommorrow's Market Share
It's sad and funny how companies with some halfdecent growth rush out and start brandishing graphs and charts where their marketeers just extend the lines from the last two years into the next 10 years or so. "Today 20% of the market, 40% next year. When we hit 100%, we just grow the market." Yes, sure. Just plug a few choice keywords, like AMD, increase and market share into the Inq search bar and see what pops out. The best one is http://www.theinquirer.net/gb/inquirer/news/2006/08/22/amd-wants-40-per-cent-of-server-market .Yes, you have to aim high but even then, as the trusty old Adamson Rust put it in another article at the time: "Let's hope AMD is not getting cocky. Remember when Nvidia got cocky?"
Somebody might also suggest, all that talk about insane growth is just designed to drive up share prices so that *some* people can cash in their share/stock options. Once there is no more milk left in the old cow, they just move on to find another udder to latch onto. What a terribly cynical view!!
Beijing to host AMD research center
Hi,I read this article in the People's Daily Online (http://english.peopledaily.com.cn/90001/6412508.html#)
and thought it might be of interest to the Enquirer.
Avid reader,
Deacon
lorencef@sbcglobal.net
Beijing to host AMD research center
U.S. microprocessor maker AMD will set up a joint research center in Beijing, as part of its commitment to the world's fastest growing PC market.
The chipmaker will collaborate with the China Electronics Standardization Institute of the Ministry of Industry and Information for the project. They will use the facility to research and develop test and evaluation technologies of integrated circuit and central processing units.
"China is one of the engines driving AMD's global evolution and today's announcements reflect our commitment to and belief in the future of China and its strong technical community," said Dirk Meyer, president and chief operating officer of AMD.
"We appreciate the strategic relationships we've developed with our customers in China and look forward to further collaboration as we bring industry-leading AMD technology to business users and consumers throughout the region."
China is now one of the fastest growing markets of the chip maker, which had told China Daily last year that it expects the nation to become its single largest market by 2009.
The company has a research and development center in Shanghai and an assembly and testing plant in Suzhou, a booming manufacturing city in Jiangsu province. In 2005, it transferred its technologies in industrial microprocessors to the Peking University, as part of its cooperation with local industries.
The chipmaker has increased its market share in China over the past years, thanks to its achievements in broadening its relationships with PC makers in China.
Over the past year, all major PC makers in China like Lenovo, Tongfang and Founder have rolled out desktops and laptops equipped with AMD processors. Even Dell, the world's second largest PC maker, decided to add AMD to it vendor list last year, rather than sticking to AMD's rival Intel.
Computer makers like IBM, HP and Dawning have also opted for the smaller chip maker for its servers, helping the company to expand in the enterprise processor market.
"We expect AMD to gain market share in China this year," said Simon Ye, an analyst with U.S. IT consultancy Gartner.
The company also donated 1 million yuan yesterday to Tsinghua University for a scholarship, trying to help nurture tech talents in the nation.
Over the past week, AMD has brought its executive operating committee to China for the first time, which shows the importance it attaches to the Chinese market, the company said.
In addition, AMD has announced that it is donating 1 million yuan in disaster relief to earthquake-hit areas in Sichuan province.
Source: China Daily
Copyright by People's Daily Online, All Rights Reserved
If AMD die
I can't realize what could happen if AMD dies.I was ATI fanboy. Now I am with DAAMIT. You know, one loves Ferrari, others Renault - and they know Ferrari is faster and looks better, but doesnt bother.
And now my Renault rides from the hell and I'm giving him Big Ups.
Very nice article.
SPIN SPIN SPIN
“With intensifying competition from rival chipmaker Intel putting the firm under pressure”
Ah the classic INQ spin rears its ugly head. You and this pathetic excuse called analysis are indicative of the bias irrelevancy you excrete. With your keen eye for objectively I would like you to answer one question as you spin your misleading ridiculous fabrications.
Had Barcelona succeeded in delivering as promised, 40% I recall, would Intel’s “pressure” be as great?
Intel did not go with the BLEEDING SOI, AMD did.
Intel went with MCM offering more product alternatives, AMD did not, and they ridiculed it.
AMD executed horribly and spent billions on a broken SOI process and a broken, untried, native quad core architecture.
AMD took on two powerful competitors, Nvidia and Intel.
You cleverly, or stupidly, completely omit Nvidia (not mentioned once) as competition.
Your ridiculous analysis maximizes Intel “intensifying competition” and minimizes AMD’s monumental failures with a mere, “delays in launching Barcelona”.
Listen, AMD has FAILED on ALL FRONTS. It wasn’t Intel’s intensifying competition. It was AMD failure to execute a viable long term successful business model, you twit.
Barcelona Failed. The ATI acquisition failed. The platform DREAM failed. Their singular offering NATIVE QUAD CORE solution has failed.
IN SHORT, AMD FAILED AND OFFERS NO COMPETITION!
INTEL IS EXECUTING!
NVIDIA IS EXECUTING!
Do you get the difference? Or maybe, in your myopic and slanted estimation, it was Intel’s monopolistic behavior? The whole article is simplistic, flawed, biased and elementary, just like its author.
SPARKS
Sparks
Intel didn't go SOI......http://www.eetimes.com/news/semi/showArticle.jhtml?articleID=10809976
dude your and ID10T and get your facts straight....
High K Dielectrics is good for one thing and SOI is for an other.... High K dielectric is good for capacitors signalling and SOI is good for insulation.... they are both different things for different purposes.. AMD and Intel use both techs.
http://en.wikipedia.org/wiki/High-k
"Silicon dioxide has been used as a gate oxide material for decades. As transistors have decreased in size, the thickness of the silicon dioxide gate dielectric has steadily decreased to increase the gate capacitance and thereby drive current and device performance. As the thickness scales below 2 nm, leakage currents due to tunneling increase drastically, leading to unwieldy power consumption and reduced device reliability. Replacing the silicon dioxide gate dielectric with a high-κ material allows increased gate capacitance without the concomitant leakage effects."
http://en.wikipedia.org/wiki/Silicon_on_Insulator
"Lower parasitic capacitance due to isolation from the bulk silicon, which improves power consumption at matched performance.
Resistance to latchup due to complete isolation of the n- and p- well structures."
You talk spin spin spin learn more about electronics and processor engineering before you speak and remove all doubt that you are a fool.
Dumb Bell
Db? Does this mean dumb bell?1st you need to do a few things.
1) Get a brain
2) READ MORE so you know what you’re talking about.
3) Go back to 1)
This is from Intel’s website:
“Silicon-on-insulator (SOI) – SOI refers to the use of a layered silicon-insulator-silicon substrate on which transistors are built, rather than a simple (bulk) silicon substrate. Some companies claim to get some performance and/or power benefits from SOI over bulk silicon. Intel’s analysis shows that such benefits, if any, are marginal, and do not justify the substantial cost increase of SOI wafers. Intel has never used, nor does it plan to use, partially depleted SOI (PD-SOI) that others are using. There is another type, however, called fully-depleted SOI (FD-SOI) that is under investigation at Intel and is not being used by any chip makers today.”
See! You read that! Now, all you need to do is get a brain to understand what you’re reading.
Hint: re read it few times maybe you won’t need a brain.
SPARKS
In With Good=New Hope
One thing litely mentioned is that AMD/ATI has hired two new engineers to head cpu developement sections & both are from IBM. this is good & strong point that dead wood is that: dead.Maybe those new rascals will incorporate All 'N Wonder....blah blah,woof woof. Those two peoples control future of Most of AMD today.
drashek
Hello sparks... where are your URL's for proof
Dude I want to believe your dumbass BS but since you do not actually provide any links or references .... stating "This is from Intel’s website:" without an actual URL and I'm should believe you... from the actual URL I used ... the title of the article and I quote "Intel to use new SOI-based transistors for low-power processors in future"and here is a little snippet (1st paragraph... read the last sentence carefully):
"SANTA CLARA, Calif. -- In a move to address the power concerns of future microprocessors, Intel Corp. here today announced a new transistor structure that marks the company's initial use of high-k dielectrics, epitaxial wafers, and a real shocker--silicon-on-insulator (SOI) technology. Until now, Intel has dismissed SOI technology for use in processor designs."
Sparks I work as a security analyst .... in other words lots and lots and LOTS of research online dude no URL then your lying otherwise prove your case by giving proof other then your pearls of wisdom.
Cheers
from your own quote...
One last thing"There is another type, however, called fully-depleted SOI (FD-SOI) that is under investigation at Intel and is not being used by any chip makers today."
Can you say 32nm ;)
Cheers Sparks