We finally ordered a bunny suit for Charlie from the Inquirer which he will be sporting at Intel’s upcoming IDF in September. As you all know AMD and Dell both announced their deal is much larger than high end four way Opteron servers. No point in saying “I told you so” :)
If anyone has some good “sign” ideas to put on the suit please drop me a line or post a comment.
The other day I wrote a blog about our ratios on the high end changing significantly and all of the sudden a bunch of “pumpers” are making claims that AMD is in trouble. Talk about reactionary crap, AMD is not in trouble – at all. In fact it’s clear that AMD has now planted the seeds which will help pull themselves into a true Duopoly position against Intel (love it or hate it guys, it is reality). It’s also great to see Intel back in the game, no pun intended; their Core 2 Duo is their messiah - but it'll be awhile before Wall Street accepts the fact that Intel will most likely change the way they do business. All indications show that Intel should become more profitable and will likely lower their overall revenues.
Such competition is good for our industry, it means that innovation is now a matter of course, and complacency = death. Now that Dell bit the big one, chances are others will follow as they accept AMD as a mainstream alternative to Intel.
Nvidia is doing incredibly well; their short term-mid term outlook is insane, especially with their upcoming Intel platforms, server chipsets, and notebook business. On the other hand, ATI has overtaken the premium performance market – I can’t remember the last time I have received so many requests for ATI discrete graphic cards! The ATI Radeon R580+ (X1950XTX) is an absolute killer card – it’s just too bad we can’t get them.
As you know AMD announced their 4X4 technology in July. Many questions have come up since the announcement about 4X4 regarding the benefits that gamers and enthusiasts will realize. With Intel Kentsfield around the corner and AMD 4X4 on the way, the concept of multi-threaded gaming is becoming an exciting reality. Of course it’s not all about multi-threaded gaming like many people assume, there are many more benefits of 4X4 that I can’t wait to get into.
I believe AMD is paving the way for a new category of enthusiast and I am writing a 4x4 article for the next issue of CPU magazine. I think AMD via 4X4 has some incredible potential to create an entirely new category as long as AMD markets it well.
…and man, oh man, there is so much more on the way, this year is going to be in-frickin-sane.
42 blogger comments:
Intel won't allow Charlie in Bunny suit into IDF, he will have to dance outside.
As I have predicted, Conroe launch time is DELL's AMD time. Conroe is the last straw that pushed DELL to AMD.
http://sharikou.blogspot.com
Also, 4x4 is a permanent solution that will pin Intel at roughly half of AMD's performance. AMD has changed the rules of the game. In servers, the standard now is 4 socket. 2P has become ultra low end. In desktop, the standard will be soon 2P. Intel miss the game.
Hello Mr. Sood,
I too think "there are two kids in town".
As Exxon & BP and other oligopolies in other markets they will find a way to do business a more "tee for two" style.
Real "pricewar" will be off some time. As you mentioned, i too think intel will smaller itself. Further i go to say that there might be a day intel appreciates to be woken up by a green arrow in the behinds. There should be some life in the company where there war "agony-inside".
The "bunny" could a sexy playboy-girl-like one with a fluffy tail. A "Voodoo" on his butt was fine. Give the bunny some (amd-) green here and there, not too much of the "IDF forbiddden" colour.
regards juergen
Rahul,
With Intel having a competitive solution for the desktop and mainstream server market, the story for the next 6-12 months will be financial, rather than technical. Dell's business is good PR for AMD, but the cut rate prices that Dell pays, the current price war, the need to overcome Intel sole-source marketing dollars, and AMD's higher production cost (due to less economies of scale, being behind in the transition to 65nm, the need to pay IBM for fab technology, and the need to pay Charter to fab) will ensure that AMD won't make a lot of money from these chips.
Both companies will experience a lot of pain, but Intel is in a much better position to weather the storm. AMD is looking at increased costs to pay for fab modifications, new fabs, and ATI. With ASPs dropping, AMD may end up losing a lot of money over the next 3 or 4 quarters.
I share your desire to see both companies succeeding in this market space, but I think AMD will have a very hard time over the next 6-12 months absent a major (and unexpected) increase in demand for processors.
Regards,
Alan
Rahul,
I think AMD will definitely become a more legitimate competitor to Intel in the traditional PC market. However, this market is doomed to become predominantly a commoditised category with a consumer PC costing $300-400 in emerging markets where the growth is. Even in mature markets this is starting to come true. That's not to say Voodoo won't have a role but your audience is niche. The growth markets will become new categories like Viiv/AMD Live, UMPC and if Intel's bet is good then WiMax. IMHO opinion AMD/ATI do not have the scale to build the software required for platforms...or move the eco-system like content players or make WiMax investments.
Hence, AMD's need to couple w/ ATI and platformize but platforms are more than just bringing CPU and chipset together.
On a seperate note, here's a suggestion for the Charlie in the Bunny suit:
"I am FAB(u)less"
Mojo
You Sharikou are a totally stupid AMD fanbois
Are you guys still located on the hooker stroll, you know the real scuzy one with all the heron addicts and freaks that make you cringe and do all the employees still spend all day gawking out the windows at them.
"INQUIRE NICELY, OR I'LL GET THE EGGS OUT!"
Rahul, Wonder if AMD can design a feature for the 4X4 motherboards so for more normal workloads, only one socket and
one graphics card are enabled..to save the watts.
OMG, stop now you god damn AMD fanboys, intel will rule the CPU market from know on and until the end of the world. AMD dosnt have a chance, intel has already moved to 65nm, and have alreadt shown a 45nm working CPU. Intel will ahve the CPU chrown for at least 10-15 years!
the current price war, the need to overcome Intel sole-source marketing dollars, and AMD's higher production cost..
You get the exact opposite picture. AMD's cost per finished CPU averages at $40. This is the average cost of all Opteron, Turion, Athlon, Sempron CPUs, well below Intel's average cost. This is why AMD has a 59% gross margin with an ASP of below $100, while Intel has a 49% gross margin at an ASP of $150. The current price war will BK Intel in 5 to 7 quarters, as Intel's prices crash by 50% and unit share suffers another 20% loss. See my price war analysis on June 2, which had accurately predicted what AMD would do to finish its job. You may read more of my analysis on similar topics at http://sharikou.blogspot.com/.
It's all about the choice. There are certain type of customers that would not go for commodity goods (Intel). Just think about it like "the market for the different people". If there were no such people, we all will be driving Toyotas, use Intel CPUs (and i740-style graphics) and depend on MicroSoft for everyday tasks. This, fortunately, is not happening.
The market would create an AMD, if there wasn't one already.
Howdy traveller,
I nticed you mentioned 4x4 and I really hope that AMD pushes it for large dev houses where developers need client server environments with 3 or 4 VMs.
I will definitely be in line for it as I have reached a plateau with 1 socket. I would definitely recommend it to anyone who wants the most powerful desktop.
I also tried to tell people that Dellwold go with AMD and slot them in the mid range Dimension area and Precision on the lowend.
I'm sure you guys are getting excellent numbers with 5000+. I can only imagine TWO of them.
I would like to see 8GB mobos though. I guess the limitation is to differentiate bewteen wksta and server but hey I'd pay a little more for the larger RAM footprint.
Anyway, it's good to see someone in the industry who understands everyone HAS TO get used to a PC duopoly (just like IBM wanted).
To the individual who implied that AMD wasn't making chips before the IBM SOI deal should really do more research.
AMD reverse engineered the 386 and built it themselves. They got Dirk Meyer from DEC and haven't looked back. There is nothing wrong with sharing the cost of new tech.
Intel should actually learn from it. They have no more special friends now that Dell has picked up AMD. I for one am glad. I work in a lot of different shops and I HATE OptiCrap with a passion. I wish Dell sold X2 3800+ and 4200+.
They totally outdo even Core 2 (it's close) with clock for clock compile times.
Fast enough is good for me with games so FX+2 (great name) will be the productivity Nirvana for devs. ANd considering that tests were done with Vista and I know MS doesn't tune for speed until RC1, you can tack on 5-10% for both FX+2 and FX.
You're right about it being exciting right now. The new XPS 700 with the Prescott water will be sick.
They just need to get some more chipset support for mobos. Retail is looking bad for COre 2 right now with only 11 mobos for it and AM2 having 52. AT least at Newegg, but theyre a good indictor of retail sales. Right now AMD has the top 5 mobos and the top 5 chips.
There is indeed a duopoly. I also heard it IMPLIED that Brisbane may have L3. Know anything about that? 1MB shared L3 would give AMD back the crown without K8L or BullDozer derivatives.
DELL is still screwing up and the AMD move will add to their downward spiral.. Dell has to get cooler designed computers, improve customer service, and get out of may stupid businesses they are in such as the printing business, the MPw, and other consumer devices..
Dell reminds me of GM and Ford. They design ugly and unreliable cars and pray somebody will buy them. They need Government vouchers (tax breaks) such as those made for SUVs. What is the sexiest American car you can think of (that is made for the public).. Give up?
sharikou,
your analysis is as good as your PhD. Oh, wait.. Your PhD defense failed.. Sorry.. My mistake
"AMD reverse engineered the 386 and built it themselves. They got Dirk Meyer from DEC and haven't looked back. There is nothing wrong with sharing the cost of new tech."
Khalif, Sharikou, et al.,
It is one thing to pump for your pimp daddy AMD, but at least get your facts right.
First, Nexgen is really the IP basis for all of AMD's successful chips. AMD bought Nexgen when all was dark and it looked like there was no hope for AMD. Turned out well for AMD, but the innovation did not come from AMD itself.
Second, the "desktop" is not going to 2P. Maybe the top of the "performance" desktop will have a 1P/2P mix, but with quad-core on the near horizon, 2P makes absolutely no sense for anything but a niche. As the number of cores per 1P scales up, increasing the number of physical sockets provides less and less of a payoff. This is basic logic.
Third, tell your sugar daddy, Rahul, to send you on a long field trip with Charlie. I'm sure with the three of you and a bunny suit you will have lots of fun and come back as grown men.
Have fun.
Sharikou,
I place large bets on AMD and Intel stock so I need to be a little more objective. Right now I'm short on AMD and don't have an Intel position although I may get back in long after next quarter's horrible results.
As far as ASPs, Intel has an enormous inventory of Pentium Ds that will either be given away or will end up in some landfill. I'm not sure what the large OEMs are paying for them (enlighten us Rahool?), but you can get some idea by looking at PriceGrabber or whatever site you use. In any event, it ain't anywhere close to $150 at this point in time. This has resulted in Intel profits falling from $2 billion per quarter to $1 billion in the latest quarter, to some negative amount in this quarter. But Intel is a very strong company and they can easily absorb the short term pain (probably a couple of losing quarters) until the Core 2s ramp up later this year. The shareholders will not be happy; they may even sack some of their senior management, but thinking the company will dissapear or even sustain serious damage is pretty far fetched.
AMD has a different problem. They have very nice processors that are really available with lots of support. In fact, if Intel weren't giving away their processors, AMD would be in very good shape. Unfortunately for AMD, they have had to drastically reduce the sell price of their processors to compete with Intel's give away program, and that will drag them into negative territory too. Unfortunately for AMD, they don't have Intel's balance sheet and things are much worse since they borrowed $2-1/2 billion to help pay for ATI.
As far as production costs are concerned, AMD has a number of additional costs relative to Intel:
- They are ~9 months behind in the transition to 65nm which allows a doubling of transistors per wafer,
- They are far behind in use of 12" wafers which reduce costs by increasing dies per wafer,
- They have to pay IBM for licenses,
- They have to pay Charter to produce dies.
These are all real dissadvantages and they cause AMD's cost per die to be significantly higher than Intel's. Actual production costs are closely guarded secrets at both companies, but if you believe AMD's cost per die is lower than Intel's, you are very much alone in that view.
That said, AMD currently has a huge (but temporary) advantage when it comes to manufacturing; They are manufacturing chips that people want rather than Pentium D chips that may end up in a landfill somewhere. Another bright spot is that they have Fab 30 cranking out wafers at an incredible rate and since this is a really good thing in a business with very high fixed costs.
By the way, 11 of my 12 computers are currently powered by AMD and I'm currently upgrading several platforms with X2s.
Regards,
Alan
Rahul,
When Dell put a Pentium Extreme Edition in their Renegade line, even though it was factory overclocked to 4.26GHz, you still called it "putting lipstick on a pig". A Pentium Extreme Edition is behind a FX-62 by about 20% in gaming as seen in Anand's Core benchmarks, but in all else it is within 5%. Now factory overclocked to 4.26Ghz I am sure it would look a lot better. Yet, at that time you carried 100% AMD procs, and never even tried to do an innovative thing by looking at a P-EE and seeing how you could get unique performance from it.
Today, Core leads in everything.... everything. Yet you only carry 50-50. Is this because of inventory issues on either side. I do believe that you should endorse only the best technology, and keep AMD down to a bare minimum percentage that forces them to accelerate their products and bring them out to market faster. Even though Intel was behind AMD in performance it was not till the end of Q1 that they were forced to accelerate their roadmap and bring out MCW faster.
AMD is in denial today about their product lagging in performance, and such things need to be done in order to force them to execute as well, otherwise complacency can set in.
Putting AMD FX in your line today, is like putting "lipstick on a pig".
"VooDooPC found all the Easter eggs!"
The whole PC market doesn't want to revert to the Intel monopoly anymore. Now it's only semi-monopoly. Vendors know it's healthy for the PC market to have choices. Even former Intel loyalists see this.
Compare the recent new AMD products announcements with numerous PR from 3rd party vendors to the older days of only a few OEMs were showing supports for AMD, it's a trend.
Intel is cutting costs, but not its FUD machine.
Between January 1999 and 2002, Intel spent about $11 billion dollars acquiring 35 companies. It took a break and then started again.
http://news.com.com/2061-11128_3-6073591.html
Intel also took Digital's StrongArm low-power processor business and renamed it Xscale.
24 July 2002
Imagination Technologies Group plc ("Imagination Technologies"), the leading provider of multimedia and communication intellectual property ("IP"), today announces that Intel Corporation ("Intel") has licensed Imagination Technologies' PowerVR MBX graphics / video core for incorporation into Intel Integrated Circuits ("ICs").
http://www.imgtec.com/News/Release/index.asp?ID=312
17 May 2005
Imagination Technologies Group plc, (“Imagination Technologies”), a leader in system-on-chip intellectual property (SoC IP), announces that Intel Corporation (“Intel”) has licensed both a member of its new generation, fully programmable PowerVR graphics and video processing IP family, codenamed Eurasia, and a next generation multi-standard video codec (video encode and decoder) IP core.
http://www.imgtec.com/News/Release/index.asp?ID=409
HP labs EPIC -> "Itanium" -> Intel
DIGITAL -> INTEL
Oct. 27, 1997 :
Key components of the agreement pending government approval:
• Intel will purchase Digital's semiconductor operations including state-of-the-art facilities in Hudson, Massachusetts as well as development operations in Jerusalem, Israel and Austin, Texas for approximately $700 million.
• Digital will retain its Alpha and Alpha-related semiconductor design teams to continue to develop future generations of Alpha. Intel will make offers of employment to other Digital semiconductor employees.
• Intel will serve as a foundry for Digital for multiple generations of Alpha microprocessors.
• Digital will develop a full line of systems based on Intel's IA-64 processor family.
• Digital will port, with technical assistance from Intel, its industry-leading Digital UNIX to Intel's IA-64.
• Intel will obtain rights to manufacture and sell other non-Alpha Digital semiconductor products.
http://www.intel.com/pressroom/archive/releases/cn102797.htm
Compaq is transferring significant [DEC] Alpha microprocessor and compiler technology, tools and resources to Intel.
NEW YORK, June 25, 2001 - Compaq Computer Corporation (NYSE: CPQ) and Intel Corporation today announced a multi-year agreement
Technology Transfer
Under the multi-year technology agreement, Compaq is transferring significant Alpha tools and engineering resources to Intel, as well as granting licenses to Compaq's Alpha microprocessor technology and compilers.
Over the next couple of years, several hundred Compaq microprocessor engineers, compiler experts and infrastructure employees will be offered employment with Intel. A portion of these engineers will remain with Compaq to complete a next-generation Alpha microprocessor development effort currently underway but will transfer to Intel as their projects are completed. Compaq's transfer of technology and resources to Intel is expected to result in an acceleration and enhancement of Intel's Itanium processor roadmap.
http://www.intel.com/pressroom/archive/releases/20010625corp.htm
*
AMD: Evolution of an Innovator
http://www.amd.com/us-en/Weblets/0,,7832_12670_12686,00.html
Hi Mr. Sood,
do you think a single or dual core variant on K8L - layout will come out with 65 nm starting end of year ?
The core enhancements should work on these also and would help a bit on the Conroe performance.
If amd integrates chipset & gpu with cpu and then bundled with a harddisk, may we see a "3 watts ipod sizes solaris server" ?
;-)
On a square meter racked in, these could make up a nice companies infrastructure.
regards juergen
It's all a bit silly.
Why is this even a discussion.
We are arguing over two American companies.
When CPU's start flowing out of China we will have the major problems.
This has resulted in Intel profits falling from $2 billion per quarter to $1 billion in the latest quarter, to some negative amount in this quarter. But Intel is a very strong company and they can easily absorb the short term pain
You are right that Intel will post losses. You are wrong that Intel is financially strong. Check out Intel's balance sheet and see why I projected that Intel will BK in 5-7 quarters. AMD's next 3 quarters will be great. Its ASP will be steady and volume will increase dramatically. Once Intel BKs, AMD will own the CPU market. So cover your short soon, your risk is unlimited.
AMD has cemented its position in the duopoly ever since it took the challenge to Intel with the Athlon and stopped being a simple clone. I believe that now more than ever, AMD is capable of holding out until it can come up with something to match the Core microarchitecture. Intel can hurt AMD's bottom line but it cannot crush it as easily.
I believe that the 4x4 will be something to look forward to. If only AMD would be able to market it as well as nVidia did with SLI. This move allows AMD to leverage Hypertransport, something Intel still does not have, in order to cope with the competition.
What will be truly interesting will happen when AMD releases the K8L quadcore and have it work in the 4x4 system. That will effectively give you an 8-core system!
No kidding it's silly. If I want this kind of debate, I can go post on AMDZone, or even better, Anandtech.
Just read the news, take it in, and then go flame the people you hate. :P
Thanks for the blog Mr. VoodooMan. I enjoy your insight. Just wish I could also enjoy my own Voodoo PC. Working in education has it's pitfalls though.
Sharikou,
You are a funny guy. Lucky for you, it won't take long to see the error of your ways. AMD's results for 2006Q3 will be described in many ways, but not even Hecter will describe them as great. Your take on the second quarter was also over the top:
Sharikou says: "Its (AMD's) ASP will be steady..."
AMD says in 2006Q2 conference call: "Second quarter sales were down from the prior quarter primarily due to the challenging pricing environment for high-volume desktop processors which negatively impacted average selling prices (ASPs)."
And this was before AMD officially axed X2 prices. With Intel starting to ramp up on Conroes, its not hard to deduce that AMD's ASPs will be under additional pressure. Such is life during a price war...
Your suggestion on the relative corporate strengths of Intel and AMD is also not well tied to reality. Links to the balance sheets of Intel and AMD are given below (Yahoo doesn't include AMD's 2006Q2 results but they are available on AMD's site):
http://finance.yahoo.com/q/bs?s=INTC
http://finance.yahoo.com/q/bs?s=AMD
Even after marking down Intel's inventory by a couple billion dollars and assuming a billion dollar loss in the upcoming quarter, Intel is still in much better position for the current price war.
For those without the time or inclination to review the balance sheets, the S&P corporate credit ratings of the two companies are given below:
Entity Local Currency Foreign Currency National Scale Type1
Intel Corp. A+/Stable/A-1+ A+/Stable/A-1+ ICR
Advanced Micro Devices Inc. B+/Watch Neg/-- B+/Watch Neg/-- ICR
Standard & Poor's investment grade ratings in order from the highest to the lowest are: AAA, AA+, AA, AA-, A+, A, A-, BBB+, BBB and BBB-. Standard & Poor's non-investment grade ratings in order from the highest to the lowest are: BB+, BB, BB-, B+, B, B-, CCC+, CCC, CCC- CC, C, D and SD.
Aside from giving you S&Ps take on the relative risks associated with the two companies, with AMD's bonds being rated a MUCH riskier investment, AMD's B+ corporate credit rating means that it pays a lot more when it borrows money. And right now, AMD is borrowing a lot of money...
Finaly, I shorted AMD at a weighted 35.60, and expect it to go down well below 20. I've lost a lot of money on some short deals, but I doubt this will be one of them.
Regards,
Alan
And this was before AMD officially axed X2 prices. With Intel starting to ramp up on Conroes, its not hard to deduce that AMD's ASPs will be under additional pressure. Such is life during a price war...
Dude, it's the mix that matters. In Q2, AMD sold 1.0 million X2s. In Q3, it will be 3.1 mil. In Q4, it will be 6.6 million. Sempron will be reduced from 5.3 million in Q2 to 4.9 million in Q4. ASP will rise.
As for why Intel will BK in 5 -7 quarters, read my blog. It's free information, and you should read it as you bet money on AMD/Intel.
http://sharikou.blogspot.com
Dr. Sharikou, whenever I read your comments, the picture of Baghdad Bob comes immediately to my mind.
Have a nice day :-).
Sharikou,
I read through your blog. You are a very smart guy but not very objective. It would be interesting to know why you are so in love with AMD or why you detest Intel so.
Your theory that Intel will go BK soon is hard to fathom. As of July 1st, Intel had approximately $6.4 billion in cash and short term investments. This has decreased from $11.3 billion on 12/31 05 mainly due to a stock repurchase plan and some paying down of long term debt. Assuming Intel maintains gross margins in the high 40s during the next two quarters, there will be some pain and a lot of griping from the shareholders, but there is no indication of any problems that couldn't be taken care of by suspending the stock buyback program and if necessary, reducing dividends. As far as reducing payroll is concerned, Intel is overdue to get rid of dead wood and this is normal and desirable for a company of Intel's size (at least from a shareholder's perspective). By the beginning of next year, Core 2s will be everywhere and we will see how the competition ensues.
As far as production costs go, you are pretty much alone in believing that AMD's costs are lower than Intel's. In the not too distant past, I've even heard high level AMD people stipulate to this. This isn't reflected in gross margins because until the middle of 2006Q2 AMD could sell everything they made, while Intel was making a lot of useles inventory. No matter how efficient you are in manufacturing, it doesn't improve you gross margin unless you sell it...
Finally, its hard to accept that even you believe AMD's ASPs will rise over the next 4 months. The X2 prices have come down a lot already and I suspect they will come down a lot more. As long as Intel is determined to provide more than 80% of the world's processors and AMD is shooting for 40%, there will be a lot of excess processors and severe pressure on processor prices. AMD's ASPs may drop by a lower percentage than Intel's, but its certainly going to be heading lower rather than higher. And overcapacity won't be limited to the desktop market...
Regards,
Alan
"I read through your blog. You are a very smart guy but not very objective. It would be interesting to know why you are so in love with AMD or why you detest Intel so."
Apparently, you have a lot of tolerance for stupidness!
IT seems you are new and did not know that Sharikou was fired from an agency which placed him at a company he despises! The reason given: System Abuse!
It is up to you to interpret what that means! Since then, he has been a professional couch potatoes with an outdated PC and all the time of the world to troll around the web..
Anyone who would take Sharikou's advice needs to remember that Sharikou has picked a side first and will formulate any sort of argument he can to support his choices, rather than looking at facts and being impartial.
Sharikou's blog is full of nonsense and visions of grandeur. He believes that top execs are listening to him.
Any psychologist worth their salt would assess him as:
1. Paranoid. When someone contradicts his thoughts he thinks they are paid Intel pumpers.
2. Obsessive-Compulsive. Sharikou must spend most or all of his day just writing in blogs about how Intel is evil. Unless he's being paid for his efforts, which I doubt anyone would do, he's doing this out of compulsion. And it's clear he's obsessed with various thoughts, which I won't even mention.
3. Delusional. Sharikou sends out 'briefings' to Intel and AMD execs and somehow feels they read them and will state in his blog how they have listened to his words. If you've read his columns for any length of time you may probably come to the conclusion that he is not a PhD either.
Anyone out there who plans on investing, please read further and do not take anything Sharikou says without investigating yourselves first. Scan through his blog first and make a solid judgement about his words to be safe before listening to anything he says or you may lose a lot of money. He is not well and is not capable of making rational decisions at this point.
Your theory that Intel will go BK soon is hard to fathom. As of July 1st, Intel had approximately $6.4 billion in cash and short term investments.
You can't just look at cash. You need substract Intel's current liabilities. After that, the net is only about $4 billion. Intel has also bout $3.5 billion long term investments, but it also has $2.6 billion long term debt. The total amount of liquid assets Intel has is about $5 billion. That's everything included, enough to sustain 7 quarters of moderate losses. IRS said Intel owes 2.5 billion taxes.
As for average cost, you can also find it from AMD and Intel's income statement. In 1Q06, AMD's total cost of production including depreciation is 533 million. It sold 13.3 million CPUs. Thus, a finished AMD CPU costs $40 on average. Intel's cost is much higher.
Actually, this is obvious. AMD's ASP (average selling price) has always been below $100, probably around $95 in 1Q06, and its gross margin was near 60%. Intel's ASP was about $150, with a 49% gross margin. AMD laughs to the bank when it sells a CPU at $100. Intel, on the other hand will post substantial loss doing that. If you look at Intel's 2Q06 results, it revenue from CPUs was 3.338+1.958 = 5.3 billion, it sold about 40 million CPUs. ASP was thus $132. At this ASP, Intel posted a operating income of $0.93 billion. Now, if you cut Intel's ASP to $100. Then Intel will post a loss of $32*40 - 0.93 = $0.3 billion. Now, because of DELL -AMD deal, Intel loses 3 million units in one quarter. The lost will be $0.67 billion. Now, add more unit losses due to Lenovo and IBM's expansion of AMD offerings.
You can try the same exercise for AMD. But keep in mind that AMD's ASP is below $100 now. Even if AMD drops ASP to $80, it can still make some nice profit and revenue growth, as its capacity is ramping up at double digit rate/q. All AMD needs to do is push Netburst against the pavement. As I suggested in the "AMD: Time to kill" article, by slashing X2 prices, AMD denies Intel the oppurtunity to unload its legacy inventory until AMD's own 65nm K8L parts ramp up. IBM and DELL signed up AMD partly because of the K8L roadmap.
Once AMD cranks out K8L on 65nm, Intel's will be pushed back by 3-5 years. By mid 2008, AMD will take 80% of the CPU market.
Sharikou,
You have fuzzy math! You are comparing a platform company that sells more products than just CPUs to a company that sells only CPUs (at least for now).
You are also assuming that Intel is sitting still waiting for AMD to get their K8L out. Do you know that the name "K8L" is a name created by Intel?!
I know you are a PhD in process engineering, electrical engineering, computer engineering, math, chemistry, psychology, finance, business, history, archeticture, ... But if you learned from history at least, you would know that Intel has lost 30% of market share before and came back with a vengence.
I do not expect Intel will repeat history by taking all the market share back. But at least, it will not die in 5-7 quarters! AMD is a great company and has great leaders, strategists, and engineers but it does not mean the market will allow it to by a monopoly like your little mind is telling you. And if your theory is any clost to being right, GE, HP, IBM, MS and others would have disappeared when they hit the $40 billion dollars in sales.
C2D, Woodcrest, and Merom are great processors and have a lot of demand (not in your fantacy land though!). And what is coming is even greater!
One thing Intel will do to servive, find smart innovators that can match AMD's talent. And trust me, they do have the Scientists and engineers. They may not have the leadership to take them there, but that may change soon.
Have your friend with the Bunny zsuit walk around witha carrot-shaped Novelty pen and/or have 2 women in sexy bunny suits escort him. Either that or have 2 IT guys with clean room Bunny suits escort him. The choice is yours Mr.Sood. Dell did not get AMD while it was hot. And now Intel got a CPU that is faster and produces les heat. AMD got to play catch up.
You have fuzzy math! You are comparing a platform company that sells more products than just CPUs to a company that sells only CPUs (at least for now).
Dude, I was assuming Intel's revenue from chipsets+MBs staying constant in my math. In fact, both should decline as Intel losing market share. So the picture will be even worse.
As for Conroe, its volume is negligible -- less than 5% for 3Q06. Intel's copy-exact is the bottleneck.
Apparently you do not know what copy exact means..
Copy exact is a strategy that is used not only at Intel, but in almost all industries. This explain the fanchise model just in case you do not know. The copy exact gets you a consistent product quality at a lower cost. Standardization of tools, labs, and processes makes it flexible to trouble shoot and stabilize processes. An engineer in Malaysia can understand and troubleshoot an issue in the USA because of the copy exact strategy.
You think AMD is not using copy exact?? Well, body, they do not have more than one fab!
Sharikou - You already missed your prediction on Intel GAAP loss in Q206. I predict AMD will post GAAP losses by Q207. And the interest on the 2.5 bln $ for the ATI acquisition will severely hit their profits. You are also conveniently ignoring that Woodcrest, Merom and Conroe are shipping for revenue in Q3 so expecting Intel's ASP to drop by 40$ is naive. Remember - consumer perception of value does not revolve around a product. It centres around perceived value of which they already have a price point in mind. They just get more for the same money. And while you downplay the 2P market in your blog (probably because you know Opteron is going to take a beating), remember this is over 90% of the server market and will drive Intel's ASPs up as they regain share...share across servers & desktops) which is inevitable.
You also oversimplify how costs are accounted but unfortunately your simple math is irrelevant in the real world. In addition, the Core 2 architecture die size is smaller compared to previous processors so once yields are normalised this means higher margins for Intel even at lower prices across the board as they dial down Netburst and up Core 2.
Nvidia will squeeze their relationship with AMD as long as they can and get into bed with Intel at the first opportunity. Do you really think they can trust the new AMD/ATI to not defend itself when they come after their market share and revenues. In the meanwhile, they will continue to try and build their own capability to have the GPU overtake and absorb the CPU but this is not an easy feat.
The war is beginning and its AMD/ATI against Intel + Nvidia. AMD will BK by end 2008 and probably be bought by nVidia if they mess up the ATI acquisition. Absorbing a company almost half your size is not easy.
AMD has more than one fab. They have two in Germany. FAB30 and FAB36. I don't believe that the newer FAB36 is running at full capacity yet though.
I wonder when Sharikou said that Intel would be bankrupt in 5-7 months?
5-7 months ago?
I bet in 5-7 months he keeps saying that Intel would be bankrupt in 5-7 months...
Alan - great post above. Good to see someone look at the business holistically. On the price drop impact to margins. Even if Intel's price drops are higher than AMD's, if they are squeezing more cost out of the equation through new process technology as well as lower cost on Core 2 then what really counts are margins. Bryant in the last qtrly update did mention that Q3 will be 48/49% but Q4 should return to the mid-50's and climb from there. If you look at Intel's roadmap, they are basically planning a series of launches...almost one a year if you count Core 2, Penryn derivative in 07, Nehalem and then Nehalem-C. I think AMD will catch up with K8L...probably pass Intel performance but after that will struggle unless they can match the pace of innovation at Intel. Intel has finally got it's head out of the dark place and figured out that this is the technology business. You can't sit on the same product (specially a bad one) for 5 years. In addition to their manufacturing, they are now using their R&D as a tactical assault weapon. I think they've figured out there is no longer a "knockout punch" in this business. Instead, they need to keep delivering a series of body blows till the competition weakens to the point of caving.
Rahul
On ATI stopping development on Intel platform, any idea whose
initiative it is, AMD or Intel?
Did Intel end the licensing deal with ATI or AMD tries to maximize its R&D efforts on AMD platform?
Thanks.
Sharikou - You already missed your prediction on Intel GAAP loss in Q206. I predict AMD will post GAAP losses by Q207. And the interest on the 2.5 bln $ for the ATI acquisition will severely hit their profits.
I was expecting Intel to book over $1 billion of impairment on goodwill due to the sale of business. However, that has been pushed to 4Q06. So instead of seeing a GAAP loss for Intel's 2Q06, you will see a far greater loss later.
AMD will be doing fine. By 2Q07, AMD should be taking close to 45% of the x86 market and profits will soar. Also, AMD will sell its SPSN stake, which worths about $0.8 billion right now. The timing of ATI acquisition was perfect. We should see K8L in 2Q07, with 60% improvement on integer performance/core and 3x FP performance/core. That will push Intel back by 3 to 5 years.
My dear Sharikou - hind sight is always 20/20. A child could call a large hit to short term operating profit for Intel if they decide to lay off 10,000 people and the Marvell/Eicon deals close all in Q4. Tell us something with real insight.
As for Intel being pushed back 3-5 years. On which front...just measurement of FP and integer performance or on their business?
As for AMD being fine. Both companies will suffer because of the price war. My prediction is AMD will take a body blow because they are starting to lose market share. With the advent of the 3 Core based products they will also lose revenue share because they are just collapsing their pricing structure on existing products. Intel however is collapsing their existing products but replacing those at high price points w/ the Core 2 line up which will buffer ASPs. Selling Spansion stake doesn't account for the interest on the 2.5 bln $ they are borrowing to buy ATI while having to make investments in fabs. I don't have the same faith that you have in AMD's ability to gain 25 points of share in 3 qtrs. Specially since momentum has begun to move in the other direction and Daryl Ostrander in his June presentation claimed they would be able to service 1/3 of the market in 2008. Also remember, the x86 market has a new player called Apple...and it's all Intel (that's about 5 points of the overall PC market). Dell will give them some share but they are losing share themselves. HP who are gaining share will now sleep with Intel to counter Dell.
This is the 5th time in 10 years or so Intel has begun a serious price war to win back. The results of the previous 4 times are there to see. AMD will bleed. And I will come back to your posts at the end of Q4 to see if AMD has 40% market share in Q4. Which means they need 20 points in 2 qtrs and the capacity to boot. Good luck on your prediction - it's an outrageous guess with no precedent in any business so you're going to need it.
Meanwhile, check this out. Opteron cluster gets beaten by a Pentium D cluster. AUthor concludes that Opteron is 2X price at same performance. This my dear sir will become the nature of the microprocessor business. It will become like the car industry in the US. Multiple segments with niched products to serve each segment. There will be no "one ring to rule them all" as we go forward. Different products will excel at different apps and consumers will have to choose the best fit for themselves:
http://www.clustermonkey.net//content/view/144/34/
I wonder when Sharikou said that Intel would be bankrupt in 5-7 months?
5-7 months ago?
I bet in 5-7 months he keeps saying that Intel would be bankrupt in 5-7 months...
Posted on the sign in front of Sharikou's Restaurant: "Free lobster and filet dinner- tomorrow"
Post a Comment