The AMD Execution Thread [2007 - 2017]

[ban25]

Sure you can. That's what restrictions on monopolists is all about. There are all kinds of things that are entirely appropriate for you to do if you *aren't* a monopolist that become illegal when you are.

I'm not sure which Microsoft vs Netscape trial you followed. The free bundling of IE wasn't found (correction: a key allegation that lead to DOJ prosecuting in the first place) to be a particularly nasty form of predatory pricing?

The courts may be able to force Intel to change its business practices, but it won't be able to force customers to buy AMD products. If AMD can't make competitive products, no court in the land will save them.

 

[Geo]

The courts may be able to force Intel to change its business practices, but it won't be able to force customers to buy AMD products. If AMD can't make competitive products, no court in the land will save them.

That observation is biting its own tail. You're saying if a court ordered Intel to charge $1M per processor that consumers wouldn't find AMD's products very competitive indeed? Of course they would.

(No, I'm not expecting a court would do such a thing; just pointing out the illogic of your observation.)

 

[Geeforcer]

Sure you can. That's what restrictions on monopolists is all about. There are all kinds of things that are entirely appropriate for you to do if you *aren't* a monopolist that become illegal when you are.

Such as? You can't accuse Intel of dumping since their margins are still very high. And again, even you you were declared a monopoly (which Intel has not been) you are still allowed to set prices. Even if Intel's legal department did not sigh off on current strategy and it did go to court, etc etc, by the time it would be resolved one way or the other (think IBM trial) the matter might not be relevant anymore.

I'm not sure which Microsoft vs Netscape trial you followed. The free bundling of IE wasn't found (correction: a key allegation that lead to DOJ prosecuting in the first place) to be a particularly nasty form of predatory pricing?

Are you saying that IE bundling was not found to be predatory pricing? Not being an ass, I am just not 100% clear what you mean.

 

[Geo]

Such as? You can't accuse Intel of dumping since their margins are still very high.

I think I'd like to see a chart of historical standards for Intel on that point. Comparing them to grocery store or diamond standards for margins isn't really going to be on point for what constitutes "still very high".

And again, even you you were declared a monopoly (which Intel has not been) you are still allowed to set prices. Even if Intel's legal department did not sigh off on current strategy and it did go to court, etc etc, by the time it would be resolved one way or the other (think IBM trial) the matter might not be relevant anymore.

Oh, there are all kinds of ways to manage this, and they are all a pain in the ass and devoutly wished to be avoided.

Are you saying that IE bundling was not found to be predatory pricing? Not being an ass, I am just not 100% clear what you mean.

It got a bit weird there at the end as to exactly what was "found", which is why I changed my original wording to reflect why DOJ went after them in the first place.

I'm not advocating an anti-trust trial against Intel, btw. Nor even suggesting it would "save" AMD even if it were later proved beyond any doubt that anti-competitive practices of the grossest and highly illegal nature were used by Intel.

I wasn't saying any of that. I was just saying I was surprised that Intel wouldn't want very much to avoid going thru such a legal process that costs hundreds of millions of dollars and typically consumes and obsesses your executive team for years on end. . . and that their current course, IMHO, increases the chances significantly that they will. The ultimate outcome of such a process is pretty nearly independant of how much you'd be willing to do to avoid going thru it, even if ultimately victorious.

 

[3dilettante]

Perhaps Intel's bean counters are feeling the pressure of Intel's large amount of inventory.

I had forgotten that concern remains about Intel's inventory. I can't find any links, but the expectation was that it would remain high.

The cost of hundreds of millions of dollars in litigation, if it ever got that high, may not be much compared to not moving billions of dollars of inventory.

I may have overestimated Intel's desire to hurt AMD, though I think moving excess product would have the desired side-effect of hurting AMD as well.


It would be closer to dumping if AMD didn't also have large amounts of inventory it's trying to sell off as well.

 

[overclocked_enthusiasm]

I'm surprised that Intel seems to be totally unconcerned by regulatory retaliation for being a monopoly engaged in a prolonged price war to drive their only major competitor out of business.

http://www.tgdaily.com/images/stories/article_images/processor_performance/20070413/data/data3.jpg

I would say AMD is the one on the offensive in this war and not Intel. Intel's part in the "war" is that they built a better tank (Core2Duo) and AMD had to retaliate with the only tool they had which was price. Even by slashing prices much more than Intel, AMD still lost 2 years worth of market share gain in one quarter...610 basis points. AMD simply got out engineered and out fabbed. I would hardly call Intel's pricing monopolistic as they simply had the high ground of technology leadership and the resulting halo effect to price stability. AMD...well you know the story.

 

[overclocked_enthusiasm]

AMD has survived previous process transitions.




Come now, the concept of predatory pricing combined with monopolistic market dominance isn't really so hard and/or novel to grasp, is it? It's not like such a finding on those grounds would in some way be a startling and ground-breaking development in the history of anti-trust.

Whether it would be enough to "save AMD" is an entirely different matter. Intel, even if found to be a monopolist, does not have a responsbility to save AMD in all and every circumstance. They do have a responsiblity to not engage in certain practices that leverages their monopolist position in order to crush AMD.

If Intel really thought they could crush AMD without a price war, as you seem to suggest is inevitable, then why are they engaged in one to the detriment of their stockholders (if it's unnecessary, then it's absolutely to the detriment of their stockholders)? That's what it comes down to. Are Intel's current margins right now significantly lower than they have been on a historic basis? Yes or no? All the rest of that is blue smoke and mirrors.

Come on Geo, 50% + gross margins is hardly predatory pricing, nor does a 10% drop in gross margins either. Yes they used to be in the 60's but competition does that to companies. The free market valued Intel and AMD parts not big bad evil Intel. You can argue that the Intel OEM relationships, rebates and anti-AMD incentives were monopolistic but the pricing and gross margins is a red herring. Intel's pricing is stable and their gross margins are again starting to rise as per their forcast. AMD simply took on a much bigger foe and Intel's superior R&D budget and manufacturing base has finally caught the lesser capitalized AMD without an answer or the ability to counter. AMD has one way out of this mess and it is another Opteron rabbit in the hat. Aside from that, the market values Intel products more highly because they have the technology lead and the halo pricing associated with it.

 

[Geo]

That's interesting. You're suggesting all this "price war" reporting is hooey? Intel isn't doing anything different than they've always done?

 

[overclocked_enthusiasm]

That's interesting. You're suggesting all this "price war" reporting is hooey? Intel isn't doing anything different than they've always done?

The proof is in the pudding. Intel's gross margins have been mildly impacted and Core2Duo pricing has remained stable and strong. Much of Intel's gross margin woes were bringing 45nm etc. ahead of schedule and having to dump P4s to clear room for all the Core2Duos etc. As Core2Duo ramps, yields and quantity of product improve Intel migrates the price down to move inventory and make room for newer inventory. They do so while still maintaining excellent healthy gross margins...albeit lower than all time highs yet still very profitable and very high.

AMD on the other hand has seen gross margins plummet from the same 60% level as Intel down to the 20's. How has Intel only dropped 10% points and AMD dropped 30-40%? Again, it is the technology leadership halo pricing that Intel enjoys and AMD does not. Take a look at the chart I linked a few posts ago...AMD was the one slashing prices like a drunken sailor not Intel. why? Because they had no chance to sell inferior products at the same prices as Intel was charging. even this didn't work and they still lost 610 basis points of market share in 90 days.

http://seekingalpha.com/article/32616

Paul Otellini

Thanks Kevin. The first quarter marked another solid period for our microprocessor business. Units were inline with seasonal patterns and ASPs held up well despite competitive pricing in the low price segments of the market.<-----AMD

The desktop segment was competitive this quarter but our pricing held firm as we shipped more Conroe processors in the first quarter than in the second half of 2006.

Andy Bryant

For the full year, manufacturing start-up costs will be concentrated in the first half as we discussed in January. These start-up costs will be lower in the third quarter than in the second, and this change should contribute at least an additional two points to gross margin in the third quarter.

Due to the lower unit costs and permanent average selling prices for microprocessors in the first and second quarters than we previously expected, we are raising the forecast for gross margin for the full year to 51%, plus or minus a few points.

The results for the first quarter combined with the outlook for the second pointed gross margin for the first half of 49%. That would imply the margin for the second half of the year would be between 52% and 54%, with the expectation that the fourth quarter is a bit higher than the third.

Uche Orji - UBS, New York

Thank you very much. Just two questions, First, if I look at your ASPs for the quarter, I'd like to say they have held up well. If I look further down the line, with the new products you have, despite that you raised your gross margin expectation for the second half. What is implied in the pricing environment within that side?

Paul Otellini

We certainly think we continue to see a competitive price environment. But if you think back, over the last year, starting mid-last year as the new products were introduced into the desktop server and mobile segments, we started to see our products differentiate themselves from the competitor.

As time passes, as we extend those products into the product stack, we think we have more than just price to compete on. So we find ourselves in a better position than we were a year ago. We think it held up pretty well in the first quarter and we raised margin percentage for the year.

Paul Otellini

Everything today is on 65 nanometers, it’s essentially all core based or derivates of the core based products that are being run through the line now.


John Lau - Jefferies & Company

Great, thank you. I just wanted to circle back to you with regards to your estimates on gross margin. It's certainly much higher for the full year. And in your thoughts about that, I was wondering if you can give us a little insight as to your assumptions of what the pricing environment would be for the second half of the year and also what the overall growth for the PC market will be by this 2007? Thank you.

Andy Bryant

I can't give do that with granularity John. What we did see was as Paul said prices held up pretty well for the first quarter a little better than we expected. We think we are starting to see some benefits of the strengths of the product line that is certainly included in our thinking as we think about the year. But I also want to point out the good news in unit costs. The focus on efficiency programs improving throughput cost and cutting overhead inside the factory is paying off. It's a combination of two. Some things that are happening at Intel right now that I think are good for us.

John Lau - Jefferies & Company

And as a follow-up, there have been some quite dramatic pre-announcements from your competition with regards to what the comments on the resale channel. Have you seen that significant amount of pick-up in market share, did you gain market share into those certain markets in Q1?

Paul Otellini

Well, I think we'll let the market share forecasters forecast the market share when all the numbers are in, because we haven't seen all the data. I will comment though that we had a very strong quarter in the channel in Q1, and that is in the face of what is normally or seasonally a down channel quarter as well. And I think that is not a matter of just pricing as was evidenced in our ASPs, particularly on desktop. The channel was very, very strong on selling Conroe, which is the Core 2 Duo desktop product.

Tim Luke - Lehman Brothers

And for the June period, your gross margin, vis-à-vis some of the prior models is moving to 48% perhaps a little lower, but the second half you are now raising. Could you just walk through the key things that are raised in the back half gross margin, vis-à-vis, what you were expecting previously? The first quarter gross margin if I am not mistaken was broadly in line excluding the reserve, with what you had expected and maybe --

Andy Bryant

No, no. I'm careful. We knew what we are doing, that's why we had a forecast.

Tim Luke - Lehman Brothers

Oh, you did?

Andy Bryant

So, what's happening is we're ahead on unit cost reductions.

Tim Luke - Lehman Brothers

Okay.

Andy Bryant

ASPs firmed up a little bit. If we had done the same math for everyone last time, we did for this time we'd have seen -- might keeps your margins actually a little elevated to where I thought it was at the beginning of the year. So, I believe we're seeing some good news as the operating of the business in both the first and second quarter. In the back half of the year, we expect to see, if you assume something along the lines of seasonal revenue pick up, fixed cost industry; you pick up some margin for that versus our prior expectations.

Like I said, we're ahead on unit costs, not just in the first quarter but that stayed ahead through the year. So, we pick up a little bit of good news for that. We think the strength of the new products is starting to help us a little bit. We picked up a little good news for that. So, especially, the business is working like you and hope it would be working right now, plus you get the uplift for the start-up costs.

 

[nicolasb]

Being a monopoly is not illegal.

The sort of thing that is illegal is using a monopoly in one sector to gain an unfair advantage over a competitor in another sector. Some people have suggested, for example, that Microsoft got into trouble over IE because of "predatory pricing". It had nothing to do with predatory pricing, it had to do with bundling. Microsoft is considered to have a monopoly in the operating system space. It was illegal for them to inextricably bundle IE with Windows, because that meant they were using their OS monopoly to gain an advantage over a competitor in the web browser market.

But giving away IE for free was, in itself, perfectly legal. The problem was that MS refused to separate IE from Windows, and even (allegedly) threatened to cut off the supply of Windows licences to OEMs who didn't include the IE icon on the desktop of their PCs. That was an abuse of monopoly. Giving IE away for nothing wasn't.

Intel using its favourable position in the CPU market to improve its performance in the CPU market is certainly not necessarily a problem.

There are ways that Intel could abuse its position in the CPU market, but aggressive price cuts is not one of them. It is alleged, for example, that Intel may have threatened to stop supplying certain OEMs with Intel chips unless they agreed to stop buying AMD chips; if it happened, that certainly would have been illegal, and an abuse of monopoly powers. But price-cutting is not anti-competitive; on the contrary, it is directly competitive and benefits the consumer. If AMD is unable to compete on price/performance ratio, that is not Intel's problem.

On top of that, there is no evidence that Intel is trying to put AMD out of business. Intel is simply trying to sell lots of Intel CPUs by selling them at an attractive price. There is nothing illegal about that. You might, perhaps, cobble together a case that Intel is misbehaving if you could show that Intel chips are currently being sold at a loss; if Intel did that, it might be difficult for Intel to argue that it was healthy competition rather than trying to eliminate a competitior. But so long as Intel is making a healthy profit per chip, it is entirely up to Intel whether it tries to sell a smaller number of chips at a higher price, or a larger number at a lower price. AMD has no grounds for complaint either way.

 

[overclocked_enthusiasm]

http://www.theinquirer.net/default.aspx?article=40419

While riddled with inaccuracies like implying that R600 is manufactured in-house, the "asset lite" or move to a fabless business model does present some interesting questions:

1. Is there enough excess capacity at Chartered and TSMC to make all of the AMD CPUs?
2. Would this fabless business design require the capacity at current AMD fabs to be used to provide sufficent capacity? If so, what company would buy these fabs and then sell the capacity back to AMD? Does that even make sense?
3. Isn't a fabless business model in the face of a competitor with their own fabs (Intel) suicide in regards to timing and ramping of new products?
4. Is there another example of a fabless business model doing well against a competitor with their own fabs?
5. Does this mean AMD would be relegated to the low end due to the lack of inhouse testing and fabs?
6. Wouldn't this slow transitions to new process nodes in comparison to Intel?

Seems like alot of difficult questions to answer for AMD to remain with the current business model they have now. While the GPU business will be largely unaffected, what does this due to the CPU business?

 

[3dilettante]

I read on Ace's that AMD's technology sharing agreement had been renegotiated.
The outside fab production cap may have been renegotiated to some undisclosed amount.

It was also pointed out, and I checked the transcript, that AMD was asked several times about limitations on how much they can outsource, and various AMD execs did not answer the question.

Pacific Crest asked specifically what the percentage cap was.
Henri P. Richard told him that they do not talk about IP agreements.

I'd like to infer that AMD managed to find a way to outsource over 20% of its capacity, either through some kind of shared venture scheme or a renegotiated cap.
For whatever reason AMD, wouldn't talk about it.

If the cap is not entirely gone, then fabless is not an answer.
Hector was the most gung-ho and he said that their options with regards to asset-light production were near infinite.

Considering that AMD barely admitted prior to Q1 that things would be a little rough, I'm guessing they're actually not as close to infinitely flexible as they'd like.

 

[INKster]

So, i guess this means the end to the planned Fab in upstate New York.
And what about the upgrading of Fab 30 to Fab 38, is it still a go ?

I wonder if turning the tables (i.e., having outside sources Fab CPU's, while turning their existing Fab's in to a similar model to that of TSMC, UMC, SMIC, Chartered, etc) wouldn't be a more intelligent move.
Most BTO foundries still use previous top of the line processes, and their customers usually don't require state of the art technology to manufacture their designs anyway...

 

[overclocked_enthusiasm]

They stated in the Q1 2007 CC that they were trimming CapEx from $2.5 billion to $2 billion for 2007. They also said the $500 million "saved" would not be used to convert Fab 30 to Fab 38 as originally intended. Something is definelty afoot with AMD's business model and it is going to be interesting to see how it plays out.

Outsourcing fabrication does not lower the risk for manufacturing issues as R520 and R600 can attest. It seems to me that Intel will be able to stay a full shrink ahead of AMD regardless of if AMD owns their own fabs or not. If that is the case, then outsourcing production seems to be a viable alternative for AMD if they can be profitable selling chips at the low-mid range. I don't see how they could have the "best" performance chips being a full shrink behind Intel unless Intel stumbles.

This of course assumes that Intel would cede the low end to AMD which seems silly to me as Intel needs to sell the lower binning chips and to keep the production lines and capacity as close to 100% full as possible. In the end, either strategy leaves AMD between a rock and a hard place. I really believe we are witnessing the fall of AMD as we know it and Intel will once again become the Microsoft of chips. Will this be good for consumers?...prob not.

 

[3dilettante]

I saw this link elsewhere:

http://biz.yahoo.com/e/070620/amd8-k.html

Can someone explain to me what this means?
I get the feeling somebody's not going to like this, but I'm not sure who.

 

[XMAN26]

I saw this link elsewhere:

http://biz.yahoo.com/e/070620/amd8-k.html

Can someone explain to me what this means?
I get the feeling somebody's not going to like this, but I'm not sure who.

I'm no banker, but it sounds like AMD got the better end of those loans by having it writen so that at a point in time, they are nolonger held in aspect against assets of the company thus making it far easier for them to reneg on them or not make payments should times get tough. If I was a bank, I would never have agreed to those terms.

 

[Arun]

At this point, it is fairly likely in my mind that AMD will sell Fab30 (aka Fab38) and let someone else convert it to 300mm after the sale. I'm curious who might be interested in buying that though, hmm...

And I'd suspect they are getting rid of 200mm equipment there much faster than anyone has been expecting, in order to get some much needed extra cash and reduce their mix of 90nm chips to improve gross margins. Whether that will be enough is another question entirely, of course.

 

[XMAN26]

At this point, it is fairly likely in my mind that AMD will sell Fab30 (aka Fab38) and let someone else convert it to 300mm after the sale. I'm curious who might be interested in buying that though, hmm...

And I'd suspect they are getting rid of 200mm equipment there much faster than anyone has been expecting, in order to get some much needed extra cash and reduce their mix of 90nm chips to improve gross margins. Whether that will be enough is another question entirely, of course.

Well IBM tops my list followed by Intel. Although, it would be nice if say TSMC bought it and place a fab closer to both ATI and Nvidia.

 

[3dilettante]

What would IBM do with the extra fab though?

They've been doing anything to use up the fab capacity they already have.

 

[silent_guy]

Well IBM tops my list followed by Intel. Although, it would be nice if say TSMC bought it and place a fab closer to both ATI and Nvidia.

Intel is very unlikely: it goes against their strategy of fab copy-and-paste.

Not sure what you mean by 'closer to...'? You don't mean physical distance, do you? Why would that be important if all the other facilities are in Asia anyway. (assembly, testing, board assembly)