AMD Announces Proposed Plan to Sell Singapore Facility

[SiliconWars]

I don't believe they are as low as 15% or as high as 30%, but that's as far as I've looked at it. There are too many variables that can cause a very large swing either way so it seems pointless to attempt to pinpoint it any better than 10%.

One big one is that we don't know if AMD's other markets will be seriously impacted by the consoles, eg graphics. We also know that AMD's graphics are getting pretty old now anyway, and prices have come down a lot in order to shift the last of the stock before Hawaii launches. Obviously this will impact margins - possibly by a lot.

The important thing for me is that there isn't much (if anything) positively impacting AMD's Q3 margins, so everything will be bringing them down, not just the consoles.

At a hunch I'd say 22-25% though.

 

[PPB]

I feel you are anything but serious.

You can try explain how "competition is better" and I can gun it down completely with little effort.

Start by considering why there aint any more x86 manufactors than there is now. Its not because they cant enter the business or reapear. In the history we had over 20 x86 manufactors. The reason is purely economic. Its simply too expensive to compete.

Why dont you think AMD uses 8B$ on R&D plus building leading edge fabs? Is it because the segment can substain it? or is it because the segment cant substain it? Do you think we would be better of with 5 AMDs or 1 Intel?

AMD have offered zero competition since 2006. Yet more money on R&D and factories havent been used before.

The segment is also with a dynamic demand and a high revenue requirement to substain the production and development. And the cost keeps increasing.

So unless you are willing to spend 5-10B$ of your tax money every year on AMD. And increasing every year in amount as well. Then you have to figure out what you want.

Do you want the illusion of competition, only to sit back with inferiour products on an inferiour process nodes. Or do you want a monopoly with the best product on the most advanced node that the planet is capable of funding in a free market? Remember, the average price wont change on the product, nomatter if you got 5 companies or 1.

I think you confuse it with a static demand segment with low entering cost. But even then, competition is far from always the best. Just look at the private US heathcare business. Multiple times more expensive and worse than state monopoly heathcare systems.

Well, you just proved you can't really explain how a potential monopoly situation would be better for the customer, just loads of non-sequiturs trying to derail from a point you can't even prove.

Told you it would be fun watching someone try to explain such thing, because you have to be delusional to think a monopoly is better than any kind of competition, doesn't matter how little it is.

Just to pick one of those fallacies: when you say that a monopoly is better if it gives you the best process node, you ignore the fact that you can't guarantee Intel would give you the best process node in a monopoly situation. Intel is giving you the best process node right now because it's the thing that separates them from their competition and makes them better than them, genius.

Obviously, I could put my dumb glasses and say "well, It WOULD be the better process node because there wouldn't be anything to compare it to!" but hey, that kind of (horribly flawed) reasoning is reserved for some other people in this thread :awe:

This forum keeps proving that even a tied duopoly wont even cut it sometimes when it comes to benefitting the customer (specially, their wallet), as brand loyalty really can skew the perception of reality and make people think, for example, 1k single GPUs are okay even if they give 30% perf for >250% price.

 

[Homeles]

Industry innovation would be higher if Intel were allowed to become a monopoly:
http://www.columbia.edu/~brg2114/files/dynduo.pdf

"In support of Schumpeter’s hypothesis, industry innovation is higher
with Intel as a monopolist."

 

[mrmt]

At a hunch I'd say 22-25% though.

And what's your sale forecast? Because if you are using the same 170MM everybody is using the other AMD divisions combined would have to have a 36.7% gross margin, which is a very significant decline. That's not what they guided at the Q&A.

 

[SiliconWars]

Industry innovation would be higher if Intel were allowed to become a monopoly:
http://www.columbia.edu/~brg2114/files/dynduo.pdf

Also

Consumers are better off under a duopoly because of lower margins: consumer surplus is higher with AMD competing against Intel than without AMD.

 

[PPB]

So let's see, trust random papers without knowing the author's motives and interests he might be defending or basic mechanisms regarding competition in capitalist economies?

Tough choice.

 

[Homeles]

Also

When the argument is this...

It benefits nobody to keep something alive that cant substain itself. And actually hurting the overall segment.

...your quotation has no relevance.

So let's see, trust random papers without knowing the author's motives and interests he might be defending or basic mechanisms regarding competition in capitalist economies?

Tough choice.

Translation: "I'm smarter than two men with Ph.Ds in economics"

I wonder which is a more likely to be correct on a particular subject, the authors of peer reviewed paper, or some kiddo on the Anandtech forums?

 

[krumme]

It could be, but then why would an AMD executive say that the console margins are in the "low double digit" levels and imply constant margins on everything else in the Q&A? I can't fathom why would the executive have that choice of words or imply the constant margins if the console chips were at 30% or even above 25%.

Just FYI there was a post where I wrote that I expected some 20% when I commented prior to read the transcript because I expected the other business to stay flat or even fall a little. Only after seeing the 6% increase is that the number makes sense, and that's why the analysts pressed AMD executives about the "mid-teens" number. Numbers around this threeshold is what I'm seeing floating around in reports.

But I'm really interested in seeing your numbers. Could you share them? What's your estimate for the console margins and where do you think the margins are falling?

I dont understand this obsession of margins for the consoles. This specific %

Firstly how is margins excactly defined?
Then why is the number so important then?

I mean as a business case factors as forecast, risk, alternatives, fit to strategy not to say calculated total profit and roi looks like major parts if not the important ??

 

[PPB]

When the argument is this...
...your quotation has no relevance.

Translation: "I'm smarter than two men with Ph.Ds in economics"

I wonder which is a more likely to be correct on a particular subject, the authors of peer reviewed paper, or some kiddo on the Anandtech forums?

No, the dude that wrote this paper thinks that.

See, we can both play the "random paper as citation of authority" game.

 

[SiliconWars]

And what's your sale forecast? Because if you are using the same 170MM everybody is using the other AMD divisions combined would have to have a 36.7% gross margin, which is a very significant decline. That's not what they guided at the Q&A.

The problem is attributing the 5-6% on all segments, when the question asked about Intel and was aimed at PC's.

What Rory said was -

Clearly, the market is moving down into the entry and mainstream price points

and

Well, I see the overall PC market in the second half growing in the mid-single digit level.

So if PC's are up 5% as Intel is guiding, graphics can still be down, and in AMD's case they will be. All other graphics revenue except the new consoles WILL be down without any doubt - that's current graphics cards with slashed prices, and old console revenue.

Like I said, because of the numbers we're talking about it can easily cause a huge swing by simply not understanding one of those numbers properly, or by wrongly interpreting comments made in the conference call.

Increasing the current Computing Solutions revenues by 5% = $883 million
Decreasing the current Graphics Segment revenues by 10% = $288 million

So the guidance of $1.415 billion minus $1.171 leaves over $244 million in new console revenue vs your $170 million. We simply don't know how badly the current graphics business will be impacted so it's still a waste of time. Note however that this would be 17% of company revenue on consoles, which is around the 20% mark they said it would be by Q4 and much more realistic than your 12% figure.

 

[mrmt]

The problem is attributing the 5-6% on all segments, when the question asked about Intel and was aimed at PC's.

That's what overall PC market is supposed to mean, isn't it?

 

[ShintaiDK]

Well, you just proved you can't really explain how a potential monopoly situation would be better for the customer, just loads of non-sequiturs trying to derail from a point you can't even prove.

Told you it would be fun watching someone try to explain such thing, because you have to be delusional to think a monopoly is better than any kind of competition, doesn't matter how little it is.

Just to pick one of those fallacies: when you say that a monopoly is better if it gives you the best process node, you ignore the fact that you can't guarantee Intel would give you the best process node in a monopoly situation. Intel is giving you the best process node right now because it's the thing that separates them from their competition and makes them better than them, genius.

Obviously, I could put my dumb glasses and say "well, It WOULD be the better process node because there wouldn't be anything to compare it to!" but hey, that kind of (horribly flawed) reasoning is reserved for some other people in this thread :awe:

This forum keeps proving that even a tied duopoly wont even cut it sometimes when it comes to benefitting the customer (specially, their wallet), as brand loyalty really can skew the perception of reality and make people think, for example, 1k single GPUs are okay even if they give 30% perf for >250% price.

I guess you couldnt argument for a competition model?

Plus you seem to lack understanding betwen volume and price. Not to mention between static and dynamic demand. And more importantly the volume and revenue needs to keep the progress going.

Intel needs to give consumers a reason to upgrade. Else they simply stay with what they got. Competition there serves no purpose, because competition is not the motivator for the company.

What competition however gives in that segment is the lack of risk taking. IA64 is an example on what happens, and that consumers actually ends up with the worse product that is now the barrier today.

You need a monopolistic control like Apple did to radically change the base to something with a higher potential.

 

[SiliconWars]

That's what overall PC market is supposed to mean, isn't it?

Yes, the overall PC market. If you read the context it's talking about Intel who do not have a graphics segment.

Clearly in AMD's case it isn't including the graphics and visual segment either, as consoles are part of that.

Vivek Arya - Bank of America - So, just one follow-up on that, Rory, I think yesterday Intel guided to about 5% or so sequential growth in their PC and server segment, is that roughly how we should be thinking about your traditional business and then whatever is left is the new opportunity?

Reply -

Well, I see the overall PC market in the second half growing in the mid-single digit level. And what I would like to do is to achieve that or exceed that so that we could slightly gain share.

So Rory is clearly saying that they hope to do the same, or slightly beat Intel's numbers, which obviously can't include graphics.

 

[mrmt]

Yes, the overall PC market. If you read the context it's talking about Intel who do not have a graphics segment.

He could have said "I expect 5% sequential growth for our computing solutions segment" in opposition to "overall PC market" and yet he didn't. That's not the kind of silly mistake an executive would make, isn't it? But either way, that's your interpretation and it's futile to continue this discussion. I didn't see the 25% number floating around but I did see below 20% margins floating here and there.

 

[SiliconWars]

He could have said "I expect 5% sequential growth for our computing solutions segment" in opposition to "overall PC market" and yet he didn't. That's not the kind of silly mistake an executive would make, isn't it? But either way, that's your interpretation and it's futile to continue this discussion. I didn't see the 25% number floating around but I did see below 20% margins floating here and there.

Rory often does that, responding like for like. It's not really a mistake, you just have to look at the whole discussion and use a bit of logic.

Clearly when talking about going up against Intel, you drop the graphics part of AMD's business, and when going against Nvidia the cpu part is dropped.

Anyway, I believe my numbers won't be far off and while 15% is just about possible I would place that at the extreme low end of probability.

 

[_Rick_]

Intel needs to give consumers a reason to upgrade.

In a monopoly, they do not.
That's the entire problem.
They can, in theory, roll down most of R&D, keep a skeleton crew working on fab efficiency, and sell us the same shit forever, at nearly arbitrary prices.

The entry cost into the market is so high, that they can essentially abuse their position to any degree they want. You want a CPU? Well, you can have one. With a built-in decay-function, so it will only last 3 years. At best.
The price? A cheap 1000 dollars for our mid range device, slightly faster than ARM's latest offering. But it runs x86, which is what you need for your software!
Pay up!


And there's another issue:
You want a graphics card? Well, you'll have to buy Intel. Because for technical reasons, nVidia graphics cards just don't work as well with our chipsets. Sorry.

...so in conclusion, monopolies are terrible, unless they are people-owned. Privately owned monopolies will strive for margin optimization, and without competition, you can apply arbitrary margins.

On the other hand, fabbing capacity is almost a natural monopoly. The costs of building a next-gen fab are such, that equal competition would essentially ruin you.
AMDs x86 future will therefore probably be just another Cyrix.

In conclusion: Let AMD die. Then smash Intel up, and take state-ownership of the fabbing capacity.

 

[Homeles]

No, the dude that wrote this paper thinks that.

See, we can both play the "random paper as citation of authority" game.

Yes, but as of yet, only one of us can link a paper from a relevant time period.

The industry has changed drastically since 2006, most notably with the beginning of the smarphone revloution that began in 2007 with the introduction of Apple's iPhone, and the tablet revolution that began in 2010 with the introduction of Apple's iPad.

You asked for an explanation of how a monopoly is better, and although I am not explicitly saying that a monopoly is better in every circumstance, I'm also don't hold the ridiculous view that nothing good can come out of a monopoly.

You don't have to believe that monopolies are better. However, your view that monopolies have zero merit to them is completely false. Extremist viewpoints, such as the one you have stated that you hold, rarely have any foundation in reality.

In a monopoly, they do not.
That's the entire problem.
They can, in theory, roll down most of R&D, keep a skeleton crew working on fab efficiency, and sell us the same shit forever, at nearly arbitrary prices.

In theory, yes. However, the way the world currently works is that eventually a competitor will surface with a product superior to the monopoly's stagnant product.

Of course, one could argue that the monopoly would have the power to shut down its competitor, which is possible athough generally illegal.

 

[AtenRa]

Competition gives you : Celeron A300, Core 2 Duo, Iris GT3e.

No competition gives you : No Overclocking bellow Core i5, marginal performance Gen over Gen, 6-cores at $600+, mainstream product releases ahead of High-End.

Im sure Monopolies are the perfect environment for every corporation but Monopolies are also the worst thing that can happen to the consumer.

 

[ShintaiDK]

Competition gives you : Celeron A300, Core 2 Duo, Iris GT3e.

No competition gives you : No Overclocking bellow Core i5, marginal performance Gen over Gen, 6-cores at $600+, mainstream product releases ahead of High-End.

Im sure Monopolies are the perfect environment for every corporation but Monopolies are also the worst thing that can happen to the consumer.

Even if there was a 1000 x86 companies. You are not gonna get some magic performance increase in legacy code.

There is hardly any demand for 6 cores. Even AMD gives up on this and goes back to 4. No amount of competition gonna change that.

No competition gives us better integrated platform. Much lower power consumption and an unhead efficiency factor.

 

[ShintaiDK]

In a monopoly, they do not.
That's the entire problem.
They can, in theory, roll down most of R&D, keep a skeleton crew working on fab efficiency, and sell us the same shit forever, at nearly arbitrary prices.

The entry cost into the market is so high, that they can essentially abuse their position to any degree they want. You want a CPU? Well, you can have one. With a built-in decay-function, so it will only last 3 years. At best.
The price? A cheap 1000 dollars for our mid range device, slightly faster than ARM's latest offering. But it runs x86, which is what you need for your software!
Pay up!


And there's another issue:
You want a graphics card? Well, you'll have to buy Intel. Because for technical reasons, nVidia graphics cards just don't work as well with our chipsets. Sorry.

...so in conclusion, monopolies are terrible, unless they are people-owned. Privately owned monopolies will strive for margin optimization, and without competition, you can apply arbitrary margins.

On the other hand, fabbing capacity is almost a natural monopoly. The costs of building a next-gen fab are such, that equal competition would essentially ruin you.
AMDs x86 future will therefore probably be just another Cyrix.

In conclusion: Let AMD die. Then smash Intel up, and take state-ownership of the fabbing capacity.

You seem to assume demand is static...

 

[AtenRa]

Even if there was a 1000 x86 companies. You are not gonna get some magic performance increase in legacy code.

We would if Intel would have released a 130W Core i7 4770K.

There is hardly any demand for 6 cores.

Obviously, thats because Intel cleverly Releases the Mainstream first and High-End two Generations later (IV-E after Haswell 1150 release). Do you remember 1366 ?? :whiste:

Even AMD gives up on this and goes back to 4. No amount of competition gonna change that.

You forgetting that Kaveri will have more computation performance (FP) than 16-core Opterons with that GCN iGPU.

No competition gives us better integrated platform. Much lower power consumption and an unhead efficiency factor.

It is competition from ARM that drives Intel to give you lower power consumption and more integration. You are too blinded to see beyond blue colors :whiste:

 

[blackened23]

When the argument is this...
...your quotation has no relevance.

Translation: "I'm smarter than two men with Ph.Ds in economics"

I wonder which is a more likely to be correct on a particular subject, the authors of peer reviewed paper, or some kiddo on the Anandtech forums?

Let me fill you in on a fact homeles. Having studied economics myself (but was not my major), economics is not based on fact, it is based on theories. If you ask 20 different economists the same question, you will get 20 different answers. Economics is not an exact science. While i'm sure you feel great that google search served you well, you can actually study a bit more - just as a random example, the President of the US actually has an advisory board of various economists so he can hear every side of the coin and every opinion. And I assure you they are all different opinions.

So we have one economist with a pretty far fetched theory. So what? You can google 20 gabillion dissertations from economists staunchly stating that monopolies are never good for any industry - and those with monopolies are heavily regulated by the Government. Is that what you want? You want intel to be run by the US government? Good luck with that.

There are normative and positive economists, and they do not agree on anything. Economic theories are as subjective as the opinions on this forum and they absolutely do not have a concensus on most topics. Furthermore, nearly every economist would argue against any monopoly type of situation being beneficial for the consumer.

This is aside from the fact that even if AMD disappeared, intel would not be a monopoly - The face of computing has changed. There is no longer an all out desktop IPC battle, the real competitors for intel are ARM SOC vendors and every indication is that Intel is facing an uphill battle here. That is intel's primary competition. Not AMD. Computing has changed - 15 years ago, 99% of consumer computing devices were intel and Microsoft based. That is hardly the case anymore, and AMD is not intel's primary competition.

 

[Idontcare]

Erh, stupid question, but is AMD not too big to fail? If AMD fails, you have Intel left as a monopoly.. Wouldnt that be a worse situation for all of us(thinking FTC aka mr scissorhands)?

The "too big to fail" policy that was rolled out to cover the banks, investment houses, domestic auto industry and the domestic airline industry had nothing to do with the consumer.

It was all about "jobs", which really meant "votes" and "taxes".

The outcome of your position though is that it basically mean no business should be allowed to go slip into obscurity (or become bankrupt) as every step along that path is leading inexorably to the next in which a monopoly solidifies.

Via should be getting government subsidies, as should AMD and Transmeta. Heck it would have been nice if a government hand out had come along and kept TI's execs fat and happy so they kept us working on x86 processors too.

The other side of that coin is the reality what a policy like that would do at Intel...no exec would be able to justify the R&D budget and capex budget to the BoD (and shareholders) knowing both the BoD and shareholders knew well in advance that the government would simply ratchet up their investment in the competition as necessary to basically forever ensure Intel was just spinning its wheels.

In that reality, Intel too would do a full stop, lay of thousands of R&D engineers, and just tread water with a lower cost structure versus sprinting full out and still only treading water with a much higher cost structure.

The incentive to sprint, or jog, versus walk while competing is the implicit expectation that your efforts are going to yield results in terms of furthering the gap between you and the competition. Take that opportunity away by having government be the equalizer and all you'll really accomplish is ensuring everyone slows down and raises prices.

Airlines have not lowered prices since they were bailed out, nor have the auto companies. It was not done for the benefit of the consumers, and whatever the motivation for bailing out an AMD (if it were on the table) the motivation would be for something else too.

But AMD does not have a strong enough lobby in Washington, so it simply isn't in the cards anyways.

 

[iCyborg]

So we have one economist with a pretty far fetched theory. So what? You can google 20 gabillion dissertations from economists staunchly stating that monopolies are never good for any industry - and those with monopolies are heavily regulated by the Government. Is that what you want? You want intel to be run by the US government? Good luck with that.

There are normative and positive economists, and they do not agree on anything. Economic theories are as subjective as the opinions on this forum and they absolutely do not have a concensus on most topics. Furthermore, nearly every economist would argue against any monopoly type of situation being beneficial for the consumer.

On one side we have two economists who approached the subject scientifically and arrived at certain conclusions. And it's not some completely abstract theorizing - they specifically looked at Intel and AMD, I linked that very article some time ago in a similar thread. On the other side we have arguments from forum posters based on nothing but hand-waving.
Yes, you may criticize that economists are often wrong, theories get refuted etc., but with just these general objections that don't attack anything specific in the study and read more like a rant against the field of economics, it's pretty much like equating tarot fortune tellers and quantum mechanics, because "scientists are also wrong all the time, and they have just theories..."

 

[VirtualLarry]

In theory, yes. However, the way the world currently works is that eventually a competitor will surface with a product superior to the monopoly's stagnant product.

Of course, one could argue that the monopoly would have the power to shut down its competitor, which is possible athough generally illegal.

But how can a competitor emerge, if the product is protected by a gov't-granted monopoly? (Patents)

The only way to truly foster competition in the tech industry, is to abolish patents.