Intel drops 2.5% marketshare to AMD

[imported_Lothar]

Originally posted by: Viditor
Marketwatch

Intel 79.1%
AMD 12.8%

Is this compared to last year(exactly a year ago) or the previous quarter(3 months ago)?
If it's the later, then that figure is meaningless.

 

[supertle55]

Originally posted by: Idontcare

Originally posted by: supertle55
Enthusiasts may make a tiny portion but these are the same people heavily influencing corporate purchases of what to buy.

I understand this is popular lore, perhaps even a myth for mythbusters to examine, but has there ever been any data purported behind this manner of claim?

Is the claim based on some cited demographic study floating around out there, or is really just a menagerie of opinion over an inflated sense one's sphere of influence in life?

(not "yours" supertle55, I am speaking to the greater collective of posters who make this style of similar posts regarding "enthusiasts are 1% of the market but we in effect make 80% of the decisions on what the world buys because everybody loves to be told what to buy and we are their overlords!")

If a company was to purchase 100, 1000, or 5000 servers, it will fall on someone's lap to spec out the machine and make a recommendation. I would find it hard to swallow to end up on a person not familiar with the technologies, computer parts, etc to make this recommendation. I would doubt a CFO or an analyst of any sort would have a strong say other then budget inputs and numbers. If the person making the recommendation is not an enthusiast, I bet he/she will have access to enthusiasts who will always be pushing/influencing the decision making process. I have nothing to back up this statement other then my personal experience in seeing how purchasing decisions with regards to computers are made in medium size to large businesses.

I think if we look back at when AMD had a "huge" market share during their Duron/Athlon era pushing out great products, I really think they made great strides because of the enthusiasts supporting them from every corner. Its now in Intel's corner b/c their chips are more friendly to enthusiasts. There seems to be a relationship don't you think of market share changes and what type of products enthusiasts prefer at any given time?

 

[geneSW]

curious... has HP improved as well? If so then this would explain the jump for AMD as AMD processors are used in HP desktops. Oh, and i'll also hint that there is a DoD agency that has a contract with HP right now and is using AMD processors (phenom II's...)

Edit: Addition.... AMD's increase seems to be over the past 6-7 months acording to the NYSE.

 

[Phynaz]

Originally posted by: Lothar

Originally posted by: Viditor
Marketwatch

Intel 79.1%
AMD 12.8%

Is this compared to last year(exactly a year ago) or the previous quarter(3 months ago)?
If it's the later, then that figure is meaningless.

It's previous quarter. Year over year things stand just about where they were 12 months ago.

 

[ShawnD1]

Originally posted by: supertle55

Originally posted by: Idontcare

Originally posted by: supertle55
Enthusiasts may make a tiny portion but these are the same people heavily influencing corporate purchases of what to buy.

I understand this is popular lore, perhaps even a myth for mythbusters to examine, but has there ever been any data purported behind this manner of claim?

If a company was to purchase 100, 1000, or 5000 servers, it will fall on someone's lap to spec out the machine and make a recommendation. I would find it hard to swallow to end up on a person not familiar with the technologies, computer parts, etc to make this recommendation. I would doubt a CFO or an analyst of any sort would have a strong say other then budget inputs and numbers. If the person making the recommendation is not an enthusiast, I bet he/she will have access to enthusiasts who will always be pushing/influencing the decision making process.

You'd be surprised how poorly managed many companies are. A guy I know worked at an engineering company that was looking to upgrade their CAD software because the AutoCAD license was about to expire. The person in charge of the decision looked at price alone without talking to anyone from engineering. Rather than getting something like AutoCAD or Microstation, they went with some bullshit no name CAD package that nobody had ever heard of. Obviously it didn't work very well and they ended up buying another AutoCAD license after wasting thousands of dollars on that other package.

Given that stuff like this happens, I wouldn't be surprised if the person picking an Intel or AMD server did it because they liked the pretty colors on the box or the sales lady gave them a BJ or whatever. This same kind of retardation is why ATI boxes had "Ruby" on the cover and Nvidia boxes had that fairy; both companies tried to cater to the furious masturbater market.

 

[Markfw]

Originally posted by: ShawnD1

Originally posted by: supertle55

Originally posted by: Idontcare

Originally posted by: supertle55
Enthusiasts may make a tiny portion but these are the same people heavily influencing corporate purchases of what to buy.

I understand this is popular lore, perhaps even a myth for mythbusters to examine, but has there ever been any data purported behind this manner of claim?

If a company was to purchase 100, 1000, or 5000 servers, it will fall on someone's lap to spec out the machine and make a recommendation. I would find it hard to swallow to end up on a person not familiar with the technologies, computer parts, etc to make this recommendation. I would doubt a CFO or an analyst of any sort would have a strong say other then budget inputs and numbers. If the person making the recommendation is not an enthusiast, I bet he/she will have access to enthusiasts who will always be pushing/influencing the decision making process.

You'd be surprised how poorly managed many companies are. A guy I know worked at an engineering company that was looking to upgrade their CAD software because the AutoCAD license was about to expire. The person in charge of the decision looked at price alone without talking to anyone from engineering. Rather than getting something like AutoCAD or Microstation, they went with some bullshit no name CAD package that nobody had ever heard of. Obviously it didn't work very well and they ended up buying another AutoCAD license after wasting thousands of dollars on that other package.

Given that stuff like this happens, I wouldn't be surprised if the person picking an Intel or AMD server did it because they liked the pretty colors on the box or the sales lady gave them a BJ or whatever. This same kind of retardation is why ATI boxes had "Ruby" on the cover and Nvidia boxes had that fairy; both companies tried to cater to the furious masturbater market.

Well, I can say that many years ago (~2002) I was talking to our purchasing department about new servers, and suggested Opterons, because at the time, they were faster, cheaper, and took 1/2 the power and 1/4 the total electricity (AC cost due to higher heat on Intel). I was told that AMD was not a true certified good platform, and Intel was the only way to go. Now 7 years later we are paying the price, as we are out of power in the data centers due to a bad choice back then. And this is not a small company ~100,000 employees.

 

[OCGuy]

I would expect a larger increase after coming off the Phenom I debacle.

 

[Regs]

AMD sells each chip for a profit margin (-cost of goods sold). The only problem is do they make enough profit off of each chip sold to cover all the operating expenses accrued in different accounting periods? The financial statements I've seen so far show that they did not, and that's no big secret, with -5.15 earnings per share.

 

[Fox5]

Originally posted by: hans007

Originally posted by: Fox5

Originally posted by: iCyborg
79.1+12.8 = 91.9%
how did you end up with 99.9?

Who's got the rest? If VIA is up to 8%, that's incredible growth for them.

well it says worldwide microprocessor revenue.

an Arm chip or a MIPs or sparc chip or power4 or something are also microprocessors.

Not sure about power, sparc, and MIPS, but Arm processors make up more than half of the world's chips for sure.

Yes...depending on which chip, AMD could probably supply as much as 40% of the market.

Would those be single core semprons? Or are you counting the out-of-date fabs as well? I doubt there's much 90nm stuff selling anymore, and 65nm production should start winding down.

 

[hans007]

Originally posted by: ShawnD1

Originally posted by: supertle55

Originally posted by: Idontcare

Originally posted by: supertle55
Enthusiasts may make a tiny portion but these are the same people heavily influencing corporate purchases of what to buy.

I understand this is popular lore, perhaps even a myth for mythbusters to examine, but has there ever been any data purported behind this manner of claim?

If a company was to purchase 100, 1000, or 5000 servers, it will fall on someone's lap to spec out the machine and make a recommendation. I would find it hard to swallow to end up on a person not familiar with the technologies, computer parts, etc to make this recommendation. I would doubt a CFO or an analyst of any sort would have a strong say other then budget inputs and numbers. If the person making the recommendation is not an enthusiast, I bet he/she will have access to enthusiasts who will always be pushing/influencing the decision making process.

You'd be surprised how poorly managed many companies are. A guy I know worked at an engineering company that was looking to upgrade their CAD software because the AutoCAD license was about to expire. The person in charge of the decision looked at price alone without talking to anyone from engineering. Rather than getting something like AutoCAD or Microstation, they went with some bullshit no name CAD package that nobody had ever heard of. Obviously it didn't work very well and they ended up buying another AutoCAD license after wasting thousands of dollars on that other package.

Given that stuff like this happens, I wouldn't be surprised if the person picking an Intel or AMD server did it because they liked the pretty colors on the box or the sales lady gave them a BJ or whatever. This same kind of retardation is why ATI boxes had "Ruby" on the cover and Nvidia boxes had that fairy; both companies tried to cater to the furious masturbater market.

I used to work at this company that needed rack servers to do some back end processing of a web service (it converted pdfs to faxes and back again). This was about 3 years ago.

We had some 2U compaq DL360s which used prescott core single core xeon 3.4ghz. So we needed some more. Granted at the time I was just a QA lead. but I said something like, if we buy DL380s with dual core opterons, you wont have to space the servers 1U apart of each other for better air flow , plus our VMware servers will be able to run with the additional cores.

Needless to say the IT staff had no idea what I was talkng about and order more DL360s with the exact same config as our old servers because this way they could just reimage the drives from the old ones.

Now I'm not talking about a server farm with say 1000s of machines. This was a rack in a closet with maybe 12 boxes in it that we needed 12 more of. So basically the DL360s cost the same as the 380s, but had half the cores and used more power and wasted rack space since they overheated. But yeah I basically gave up. There was a lot of that type of stupidity (or worse) so I quit after 5 months of working there. The accountants and idiots ran the place. i.e. i got turned down for requesting an optical mouse since it cost $10 and I had a "perfectly good" mouse already.

 

[hans007]

Originally posted by: Fox5

Originally posted by: hans007

Originally posted by: Fox5

Originally posted by: iCyborg
79.1+12.8 = 91.9%
how did you end up with 99.9?

Who's got the rest? If VIA is up to 8%, that's incredible growth for them.

well it says worldwide microprocessor revenue.

an Arm chip or a MIPs or sparc chip or power4 or something are also microprocessors.

Not sure about power, sparc, and MIPS, but Arm processors make up more than half of the world's chips for sure.

Yes...depending on which chip, AMD could probably supply as much as 40% of the market.

Would those be single core semprons? Or are you counting the out-of-date fabs as well? I doubt there's much 90nm stuff selling anymore, and 65nm production should start winding down.

Yeah but its by revenue, and ARM chips do not sell for $100 a unit. Or in intels case $200 a unit ASP.

 

[Hulk]

This signals the beginning of the end of Intel!!!
















(yeah right)

 

[ShawnD1]

Originally posted by: Hulk
This signals the beginning of the end of Intel!!!

For a second I thought this was slashdot

headline: someone actually likes Linux
first post: OMG MICROSUCK IS SO FUCKED

 

[Dadofamunky]

Originally posted by: supertle55
Enthusiasts may make a tiny portion but these are the same people heavily influencing corporate purchases of what to buy.

Boy, not in the companies where I work!

 

[Viditor]

Originally posted by: zsdersw

Originally posted by: Viditor
Yes...depending on which chip, AMD could probably supply as much as 40% of the market.

.. says the resident AMD cheerleader and apologist.

So, in this hypothetical scenario, if the planets are in alignment and the sky is partly cloudy instead of partly sunny, AMD could supply as much as 40% of the market. Nice weasel-word-laden response, Mr. V.

OIC...the rabid Intelista is poking his head up...
Do you just pull these ideas out of your backside?
Let's do the math we've all seen a thousand times before...

Shanghai...at 243mm2, that's ~256 candidate dice per wafer. At a minimal 80% yield that's 204 dice per wafer
If all of AMD's chips were that large, they could produce (based on the capacity of 45k wspm for all their Dresden Fab space, and if GF made only AMD chips) 9,180,000 shanghai chips per month. Worldwide shipments are around 80 million/quarter...

Now what percentage of demand do you suppose will be for those 243mm2 chips, and what percentage for the 126mm2 X2, or the 117mm2 Athlon II???

As I said (and have now shown with the math), 40% is not a problem...

 

[Viditor]

Originally posted by: Regs
AMD sells each chip for a profit margin (-cost of goods sold). The only problem is do they make enough profit off of each chip sold to cover all the operating expenses accrued in different accounting periods? The financial statements I've seen so far show that they did not, and that's no big secret, with -5.15 earnings per share.

Going backward, the answer is no...but the vast majority of those expenses are going to be gone now (remember that they no longer own the Fabs). So, going forward, those same profits should show some very nice returns...we shall see.

 

[hans007]

Originally posted by: Viditor

Originally posted by: Regs
AMD sells each chip for a profit margin (-cost of goods sold). The only problem is do they make enough profit off of each chip sold to cover all the operating expenses accrued in different accounting periods? The financial statements I've seen so far show that they did not, and that's no big secret, with -5.15 earnings per share.

Going backward, the answer is no...but the vast majority of those expenses are going to be gone now (remember that they no longer own the Fabs). So, going forward, those same profits should show some very nice returns...we shall see.

they probably make money on an operating basis now.

so obviously they make money on top of the materials cost , so if it costs them $20 in raw materials and actual packaging and testing they are making money.

what they are losing money on is things like debt payments for bonds they put out to build the fabs, buy the equipment, and most of all the value of ATI.

r&d and advertising and all those other things are also part of the equation but I think its mostly the debt payments and restructuring of the ATI loss of "goodwill". Basically they "lose" money because they paid 5 billion for ATI. Because ATI has "lost" value on their books, they have to write that off (and they are doing it a little bit per quarter). I guess its sort of like if AMD had bought a bunch of ATI stock for its personal stock market account and its value fell (and thats basically what they did except they bought all of it)

I mean if the "value" of ATI goes up in the future (Say nvidia dies, and no one has any choice) I would assume they could write the value back in and "make" money on ATI.

THe banks have done pretty much the same thing with mortgage backed securities. In theory if houses quadrupled next year the bonds they hold would be worth more and they could "make" money on paper. Of course they wouldnt actually have any more money. Or less money. It is just based on some previous purchase because it is the booked value of that stuff (if they mark the mortgages up to some number, and dont sell them they dont take in any actual cash right, which is sort of why we have all this discussion about "marked to market" on mortgage securities. AMD loses becuase it marked ATI down to its market value basically). ATI is sort of like a toxic asset because it may never be "worth" on paper what AMD paid. buying ATI is what causes the losses (or maybe buying fabs, that didnt get used to max capacity), but i think everyone here will agree that they HAD to buy new fabs and ATI to stay in the game to begin with. Maybe they bought for too much or the economy sucked and it devalued etc, but yeah thats just tough luck for AMD I guess.

I am not a financial analyst or something so I may have gotten some of that wrong as far as the definition of "operating profit" but I think it is mostly correct. The loss is on net income, and a lot of it is because of depreciation of ati value. they have written down something like half of the 2.5 billion dollar value of ATI already so I think they probably wont have to write it down a lot more in the future (i'm not sure what ATI would be worth as a company on its own, but nvidia is worth 5 billion or so if you subtract their cash holdings and AMD I think currently values ATI at something like 2.5-3 billion of the original 5 billion)

 

[hans007]

Originally posted by: Viditor

Originally posted by: Regs
AMD sells each chip for a profit margin (-cost of goods sold). The only problem is do they make enough profit off of each chip sold to cover all the operating expenses accrued in different accounting periods? The financial statements I've seen so far show that they did not, and that's no big secret, with -5.15 earnings per share.

Going backward, the answer is no...but the vast majority of those expenses are going to be gone now (remember that they no longer own the Fabs). So, going forward, those same profits should show some very nice returns...we shall see.

That really remains to be seen.

Just because AMD itself doesn't have to pay for the interest on the fab debt anymore, or say the cost of upgrading the fabs etc doesn't mean someone doesnt have to pay for them.

Global foundaries will have to be profitable at some point too , just the arab investment basically allowed them access to cheap financing for now. Global foundaries will obviously charge AMD for whatever it costs them to maintain the fabs and pay for them, so AMD's cost of production will increase.

The main "value' of spinning off global foundaries is it allowed the arab investment (Which allowed the extra money to flow in for say the NY fab). In the future this would presumeably allow global foundaries to have other customers to better use the fab investment (which AMD really didn't want to have to bother with, though in theory they could have and eventually were going to use anyway by say fabbing ATI chipsets and igps in house).

That said being fabless doesn't automatically make them any more profitable. Sandisk is fabless but they have a higher cost of production and I believe actually lose money on an operating basis (so sandisk loses money on every piece of flash they sell, at least they did for a while)

 

[Viditor]

Originally posted by: hans007

Originally posted by: Viditor

Originally posted by: Regs
AMD sells each chip for a profit margin (-cost of goods sold). The only problem is do they make enough profit off of each chip sold to cover all the operating expenses accrued in different accounting periods? The financial statements I've seen so far show that they did not, and that's no big secret, with -5.15 earnings per share.

Going backward, the answer is no...but the vast majority of those expenses are going to be gone now (remember that they no longer own the Fabs). So, going forward, those same profits should show some very nice returns...we shall see.

That really remains to be seen.

Just because AMD itself doesn't have to pay for the interest on the fab debt anymore, or say the cost of upgrading the fabs etc doesn't mean someone doesnt have to pay for them.

Global foundaries will have to be profitable at some point too , just the arab investment basically allowed them access to cheap financing for now. Global foundaries will obviously charge AMD for whatever it costs them to maintain the fabs and pay for them, so AMD's cost of production will increase.

The main "value' of spinning off global foundaries is it allowed the arab investment (Which allowed the extra money to flow in for say the NY fab). In the future this would presumeably allow global foundaries to have other customers to better use the fab investment (which AMD really didn't want to have to bother with, though in theory they could have and eventually were going to use anyway by say fabbing ATI chipsets and igps in house).

That said being fabless doesn't automatically make them any more profitable. Sandisk is fabless but they have a higher cost of production and I believe actually lose money on an operating basis (so sandisk loses money on every piece of flash they sell, at least they did for a while)

But it also means that all of those salaries and expenses are off of AMD's balance sheets as well.
It means that AMD no longer has to rely on strictly AMD products to keep the Fab busy, they now can be far more flexible.

 

[zsdersw]

Originally posted by: Viditor
Let's do the math we've all seen a thousand times before...

Shanghai...at 243mm2, that's ~256 candidate dice per wafer. At a minimal 80% yield that's 204 dice per wafer
If all of AMD's chips were that large, they could produce (based on the capacity of 45k wspm for all their Dresden Fab space, and if GF made only AMD chips) 9,180,000 shanghai chips per month. Worldwide shipments are around 80 million/quarter...

Now what percentage of demand do you suppose will be for those 243mm2 chips, and what percentage for the 126mm2 X2, or the 117mm2 Athlon II???

As I said (and have now shown with the math), 40% is not a problem...

These things don't occur in a vacuum. What you're talking about is. The estimate is meaningless.

 

[Viditor]

Originally posted by: zsdersw

Originally posted by: Viditor
Let's do the math we've all seen a thousand times before...

Shanghai...at 243mm2, that's ~256 candidate dice per wafer. At a minimal 80% yield that's 204 dice per wafer
If all of AMD's chips were that large, they could produce (based on the capacity of 45k wspm for all their Dresden Fab space, and if GF made only AMD chips) 9,180,000 shanghai chips per month. Worldwide shipments are around 80 million/quarter...

Now what percentage of demand do you suppose will be for those 243mm2 chips, and what percentage for the 126mm2 X2, or the 117mm2 Athlon II???

As I said (and have now shown with the math), 40% is not a problem...

These things don't occur in a vacuum. What you're talking about is. The estimate is meaningless.

Your reply reminds me of the 3 monkey fable from Aesop...
This pesky math stuff really is a worry, isn't it?

 

[Denithor]

Good explanation there, hans.

So AMD needs to either cut overhead drastically so their operating profit is not exceeded by other expenses or else they've gotta sell a crapload more chips to cover these fixed costs.

In theory there would be a point where operating profit was adequate to balance the fixed negatives, above which they would be truly profitable. I say in theory because the "fixed" costs change over time also (staff costs, debt payments, etc).

Unfortunately they have not been at that point in quite some time. And because their chips aren't as fast as Intel's they have to sell for much lower margins than are healthy in order to remain competitive with Intel's offering. So unless there's a rabbit in their hat (or somewhere, anyway) this trend looks set to continue for a while.

 

[zsdersw]

Originally posted by: Viditor
Your reply reminds me of the 3 monkey fable from Aesop...
This pesky math stuff really is a worry, isn't it?

Not at all. I, AMD, or anyone else could do a lot of things based on a pure capacity perspective. On paper is quite a bit different from in practice.

 

[Idontcare]

Originally posted by: bgeh

Originally posted by: Idontcare
I had no idea AMD's marketshare had fallen so badly, down to 10%!?

Nice to read AMD increased marketshare nearly 25% to 12.8% now.

That's got to breath some hope and morale into the employees to see 25% more product moving.

It's revenue share, not marketshare:

http://www.techreport.com/discussions.x/17042

Just saw you posted a reply to my post, thanks for clarifying that for me, yeah that makes a TON more sense to me now given that AMD's ASPs are likely to be considerably less than Intel's so their percentage of the revenue share is going to be much lower than their percentage of marketshare.

Originally posted by: zsdersw

Originally posted by: Viditor
Let's do the math we've all seen a thousand times before...

Shanghai...at 243mm2, that's ~256 candidate dice per wafer. At a minimal 80% yield that's 204 dice per wafer
If all of AMD's chips were that large, they could produce (based on the capacity of 45k wspm for all their Dresden Fab space, and if GF made only AMD chips) 9,180,000 shanghai chips per month. Worldwide shipments are around 80 million/quarter...

Now what percentage of demand do you suppose will be for those 243mm2 chips, and what percentage for the 126mm2 X2, or the 117mm2 Athlon II???

As I said (and have now shown with the math), 40% is not a problem...

These things don't occur in a vacuum. What you're talking about is. The estimate is meaningless.

Its not really meaningless is it? It provides a valid upper-boundary estimate of what AMD's marketshare could be.

 

[Phynaz]

Originally posted by: Viditor

Originally posted by: Regs
AMD sells each chip for a profit margin (-cost of goods sold). The only problem is do they make enough profit off of each chip sold to cover all the operating expenses accrued in different accounting periods? The financial statements I've seen so far show that they did not, and that's no big secret, with -5.15 earnings per share.

Going backward, the answer is no...but the vast majority of those expenses are going to be gone now (remember that they no longer own the Fabs). So, going forward, those same profits should show some very nice returns...we shall see.

You're failing to read the agreement again. AMD is on the hook for a portion of the losses sustained by GF.