I want to move past the earnings announcements quickly cos they're stale news and I've taken too long to get to this. Hence, just a quick note on AMD:
1. Overall, their quarter came in on track more or less as I'd predicted including the fact they'd have a better quarter of it than Intel financially. The market share discussion also sees some light with the following assessment from Mercury Research - that both gained share from Via but Intel gained 3 points of share while AMD gained 1 point. The key here is AMD's leaps in market share are abating.
Q3 06 Market Share
2. The inventory issue (up by 15%) is worrying and my assessment is that this build up is really a result of Dell not being able to sell everything that was built for them...or just having the wrong SKUs.
3. Margins took a substantial beating down ~5.5%. A few things happened here. First, AMD's growth in servers slowed down relative to the pace they've been managing past few quarters due to Woodcrest. Those margins were allowing them to sell client parts (desktop specifically) at a large discount to gain share. Second, Intel competitive pricing really hit them and they had to take big price drops in the desktop segment which is where they have strength after server. Third, they are chasing low margin deals like Dell to secure entry into new customers like Dell and the PRC OEMs.
4. Unit volume growth was high but overall market share gain was just 1%. Intel took 3% - all of it from Via. I'd also predicted that they would do everything to keep their factories full and it looks like they did.
5. They are clearly short-changing the channel on supply and price competitiveness and this is a chink in their armour that will hurt them over the next few quarters.
Looking forward, AMD didn't provide much in the way of guidance. My thoughts:
- It is crucial for them to win back Opeteron share where they are losing to Woodcrest. I suspect they think Dell will help them do that but you cannot have Woodcrest grow to 40% of the DP market by unit and assume it's all the low end of the market from a system pricing POV.
- They should hold desktop pricing even if it means losing share and hope pricing stabilizes as Intel & AMD have said it might.
- Mobile - they should be willing to drop price in this segment to gain entry into new customers before Intel launches Santa Rosa. Frankly, even though AMD claims a 50% sequantial growth in mobile QoQ, considering the small base they have this is probably not as impressive as it sounds when you think the mobile market is growing at 20-30%. If they hope to dent Intel's stranglehold on this high margin segment they must be willing to make some trade offs.
Overall, looking into Q4 I think AMD still has it's job cut out. They will reap the benefits on cost and hence margins moving to 65nm but will be negated by their need to sustain market share. I would not be surprised if Intel extended it's Oct 22 price drop beyond low end desktop SKUs if required to keep AMD on the defensive. They need to gain market share because adjusting inventory to customer needs is easier said than done so that optimization they talk about may or may not happen. The incremental 60 million worth inventory accumulated this qtr if they've built the wrong product is going to either have to be sold or written off. Either way, it's going to hurt. Operating cash is important because they now have an ATI acquisition to pay debt on. AMD has a dilemma pulling them in two opposite directions. I suspect in Q4 they will actually cede share to hold pricing and margins. While everyone is focussed on AMD starting 65nm shipments, Intel too is increasing their 65nm and reducing 90nm every day so the battle for cost efficiency is tough...specially since the Core products have smaller die sizes.
With nothing left of Via, next quarter will truly tell who is gaining share and who is losing it. I still predict AMD will lose share and Intel will gain it. Remember, 3% of the remaining market is more than 1% in absolute volume terms. Momentum is swinging back to Intel and Q4 will show us that.
Subscribe to:
Post Comments (Atom)
17 comments:
Source for the ASP decline number? It is counter-intuitive that Intel loses 3% margin and AMD ~5.5% and yet Intel's ASP decline would be higher than AMD's. I think you need to substantiate this number.
Suggest you read the earnings call transcript & specifically the analyst Q&A. The inventory build is a result of a mis-match between the products they built and the products their customers wanted.
ASPs - traditionally Intel's CPU ASPs (which is not indicative of margin due to the diversification beyond CPU) have been in the 130-140$ range while AMD have been in the 85-95$ range. However, it's hard to tell this every qtr and the data I have is from an analyst report late last year. So just use this to get a sense of scale. This delta is not a function of the market today but how they were pegged relatively in the past. Also, Intel has a signficant premium due to their brand. Obviously, neither of them wants to change their business model to try and tank their own ASP dramatically so product and pricing planning always trys to hold or increase ASP - the exception being in price war situations. But the hope is stabilization and increase once the price war is over.
On the margin decline - if you notice, I've been mostly using QoQ as a reference for my analysis. Reason being I'm focussing on where the momentum is heading. YoY is great theory...but the true market dynamics are best assessed by QoQ and then using some judgement to look at factors that may not be consistent with the same period last year. Example - pricing environment, new product launch, Vista launch, etc.
I agree, Intel has a tougher time fixing margins - specially due to the size of their flash business. Which is why I continue to believe that business will be sold or JVed off to get it off the books. Having said that, Intel does enjoy certain cost advantages because of their manufacturing scale as well as being a process ahead. Regardless of the numbers Sharikou cooks up (based on inaccurate data which I have pointed out to him) - scale allows you manufacturing efficiencies. For example, by the time Intel's chipsets will be on 90nm, most of that machinery will have been fully depreciated having been used to manufacture CPUs first. Or the kind of pricing Intel can command from suppliers due to their size cannot be matched by anyone...not even Samsung who is #2. Finally, the significantly smaller Core die sizes across products will start helping their margins a little in Q4 (mix assumes ranges from 25% d/t to 40-60% mob/server) but much more as they ramp this product in 2007.
You are absolutely right, Intel has been using their CPU profits to fund other businesses. Specifically Barrett in the 90's spent 10 billion $s on a communications shopping spree that resulted in nothing. However, they have started vesting those businesses. The big monkey on their back I think is flash. They also have chipsets and motherboards which drive ASPs and margins down. So this is not an apples to apples comparision of margins. But the underlying fact is QoQ AMD's margins took a bigger hit than Intel's. This is a big cause for concern considering the significant extraneous factor was pricing. I'm trying to figure out whether there is a way to calculate blended ASPs for both. Will let you know if I succeed.
I don't know how soon Intel can sell their flash business seeing as they're trying to push the idea of a hybrid flash drive. While the business itself isn't doing well now, they have to make the decision of whether or not to give up in the middle of a big push for this new type of product.
While QoQ is able to consider certain facts that YoY is able to consider, QoQ cannot consider seasonal demand or short term discrepancies. YoY cannot consider non-seasonal demand changes or short term discrepancies either. In reality, you have to use both and a bit of analysis. That's why Intel's margins look worse. Pricing environments and the like are important factors, because they drastically effect the resources either company will have to do business in the future, which is what all this analysis is about.
To explain why Intel's margins would look worse then, you'd consider how little they rose, to how much higher they were last year, and consider that overall, bad. Even then, this isn't perfectly accurate, but I'd say with their current situation, it isn't easy to argue that it's very good.
Flash - I don't expect them to dump the idea of Robson. But they don't need to have a flash business to do that. Just a tie up with someone who does. In this case, they may divest the flash biz to Micron and take a JV stake in there. Hence, if Micron's qtr is not good, it doesn't reflect on their core CPU business results. Also, this forces the flash guys to not expect the profitable businesses to bale them out.
QoQ vs YoY - I agree. I made a similar comment here:
http://sharikou180.blogspot.com/2006/10/comparision-points.html
I'm not claiming Intel's margins are good. But, when Hector or Otellini think every day about how to shape their business, they're not thinking about what it was like a year ago. They're looking at what's happening in the market now. So, when I'm trying to call what will happen next, I'm not hung up on what happened a year ago. I'm more focussed on recent events to make my analysis. Based on what I see happening now, I believe AMD is taking a bigger hit as a result of low prices to Dell along with the general price war. They're also not growing as fast as they were previously so one has to wonder whether the fact they gained only a point of share and none of it from Intel could be a cause for the bloated inventory.
both gained share from Via but Intel gained 3 points of share while AMD gained 1 point.
Sorry, but your interpretation is way off. VIA had a bump in volume last quarter because of a large stock reduction of EOL products. It is completely false to describe this as VIA's taking share from Intel and AMD or Intel's and AMD's taking it back. Also, the percentage displacement of share was proportionate so, again, no change. Therefore no gain for Intel.
The key here is AMD's leaps in market share are abating.Q3 06 Market Share
Not really. The first key is to realize that AMD held onto its share in an expanding market which is good news. If AMD holds the same share in Q4 06 this will be bad news for Intel. Ideally for Intel they would gain share in Q4, Q1, and Q2. However, I doubt this will happen.
2. The inventory issue (up by 15%) is worrying and my assessment is that this build up is really a result of Dell not being able to sell everything
Sorry. This buildup is normal for AMD. If anything the buildup is smaller than normal. I don't understand how you end up interpreting this backwards. Too small of a buildup by AMD will allow Intel to gain more volume. This is pretty straightforward. Intel should be hoping that AMD's inventory is too small, not too large.
3. Margins took a substantial beating down ~5.5%.
Which is still higher than Intel. In fact, AMD's overall drop from recent high was 11% versus Intel's 20%. I would say that Intel took the bigger beating.
First, AMD's growth in servers slowed down relative to the pace they've been managing past few quarters due to Woodcrest.
True.
Second, Intel competitive pricing really hit them
Third, they are chasing low margin deals like Dell to secure entry into new customers like Dell and the PRC OEMs.
Really? Then why are Intel's margins down? The truth is that Intel had no change in revenue share nor any change in volume share which means no change in ASP relative to AMD. If Intel gained in servers then they had to lose just as much elsewhere. If AMD lost in servers; they had to gain just as much elsewhere. You seemed to have completely missed this point. Your theory about low price points from Dell is just fantasy. In reality, Intel is the loser in ASP on the desktop.
4. Unit volume growth was high but overall market share gain was just 1%. Intel took 3% - all of it from Via.
Sorry, but wrong. Intel had no volume share gain. Zero. AMD and Intel maintained the same volume share. Saying that Intel gained 2% more than AMD is very poor math.
I'd also predicted that they would do everything to keep their factories full and it looks like they did.
Then what was Intel doing? Remember Intel had no change in revenue share nor any change in volume share. If you are suggesting that AMD somehow lost then Intel had to lose as well. Strangely though you seem to be able to think of reasons why AMD has lost but can't think of reasons for Intel even when the performance is the same.
5. They are clearly short-changing the channel on supply and price competitiveness and this is a chink in their armour that will hurt them over the next few quarters.
Let me guess. If AMD's capacity was greater than demand you would say that demand was falling for AMD or that AMD would be hurt by overcapacity. Basically, you seem to be saying that you can think of a downside no matter what AMD does. Why do you have trouble seeing Intel the same way?
I'm sorry but your assessment of AMD's position with desktop, server, and mobile is way off.
The toughest thing for AMD right now is holding server share. However, Intel is going to face much more desktop and mobile pressure than they wanted.
moving to 65nm but will be negated by their need to sustain market share.
Not really. AMD's costs will drop twice as much as Intel's. Intel is going to have trouble matching AMD on price without holding its own margins down.
I would not be surprised if Intel extended it's Oct 22 price drop beyond low end desktop SKUs if required to keep AMD on the defensive.
Which won't work.
The incremental 60 million worth inventory accumulated this qtr if they've built the wrong product is going to either have to be sold or written off.
This is incredible. Intel has the largest overstock in its history, completely at odds with its normal patterns and you brush that aside. However, when AMD builds inventory as it normally does in third quarter you can't wait to start shouting gloom and doom. This seems to be more than a little bias.
I suspect in Q4 they will actually cede share to hold pricing and margins.
Sorry, but this is wrong again. AMD's margin will come up 2% even if prices stay the same. AMD's volume will definitely rise. What Intel has to hope for is that the market in Q4 expands more than what AMD can match. Because if AMD hits capacity then anything over that will go to Intel.
While everyone is focussed on AMD starting 65nm shipments,
You seem to have forgotten that Intel is already on 300mm while AMD is still ramping. This too reduces AMD's costs.
With nothing left of Via,
Huh? Via simply dropped back to their normal volume. Oh, that's right, you want to believe that Intel somehow stole share from a tough competitor. Sorry, the truth is that Q2 was unusual for VIA and did not represent their true volume. This is their normal volume.
next quarter will truly tell who is gaining share and who is losing it.
Actually, it won't. If the market volume increases more than AMD's capacity then Intel will definitely have an increase in volume share. However, this is only temporary. Okay, let me put this a different way. It is easiest for AMD to increase share in a falling market and easiest for Intel to increase share in a rising market. However, neither of these are actual gains or losses. Q1 should be a better indication. By Q1 Intel's C2D capacity will be up and yield will be good. In Q2, Barcelona will hit which will probably bump AMD's server share a little. However, the real match should be Q3 07 when AMD and Intel should be pretty closely matched on their full lines.
I still predict AMD will lose share and Intel will gain it.
I think AMD may slip a bit more in servers, probably hold on the desktop, and gain in mobile.
Remember, 3% of the remaining market is more than 1% in absolute volume terms.
An an arithmetic grasping at straws.
Momentum is swinging back to Intel and Q4 will show us that.
It hasn't swung yet. BTW, I should probably mention that temporary gains in volume by driving price down are elastic. This means that if Intel is truly relying on this they will have another crunch in Q2 07.
They're also not growing as fast as they were previously
Actually, they are. They grew quite a bit in Q3.
I understand that you tend to be somewhat blind about this in terms of the last two years. AMD's 2006 revenues should be double what they were in 2004 while Intel's will roughly match what they were in 2004. I realize that this makes Intel look bad that they slipped two years in revenue which is why you want to ignore it.
However, even looking at Q2 to now how could you miss the fact that the volume growth was big? AMD is indeed growing. And, much more amazingly is that they managed to grow and hold position in spite of Intel's greater capacity. That is impressive. The area they did not grow in was revenue but then neither did Intel.
so one has to wonder whether the fact they gained only a point of share and none of it from Intel could be a cause for the bloated inventory.
No. There was no gain from VIA. And, again, AMD does not have bloated inventory. If anything, AMD's inventory may be low. I honestly don't understand how you misinterpret these things.
Let me see if I can sum this up for you correctly:
Good for Intel:
gained server share
gained desktop share
stopped loss of revenue share
stopped loss of volume share
Bad for Intel:
revenues down to 2004 levels
loss in mobile share
no overall gain in either volume or revenue share
Good for AMD:
held revenue and volume share in rising market
gained in mobile share
revenues doubled since 2004
Bad for AMD:
loss in server share
no further gains in volume or revenue share
Potential problems:
With increasing capacity from both Intel and AMD the market will have to expand to absorb the capacity. This will be partly absorbed by dual parts with Kentsfield and 4X4. However, if the market does not absorb the capacity then this will likely lead to further price pressure. In a price pressure market, cost efficiency is more important than benchmarking performance. AMD's cost will drop faster than Intel's however Intel could divest more.
A price driven market would actually favor AMD in terms of costs but could favor Intel if volume rises beyond AMD's capacity. A reduced market should favor Intel in terms of chip performance but could also hurt Intel in terms of a volume ceiling. This is the reality and it isn't simple.
ashenman said...
"..AMD also can count on ATI's consumer electronics segment now that they've completed the merger..."
Yeah, they can definitely count on ATI's money losing business now that they've lost 70% of their revenue coming from intel based systems. AMD cannot supplement what they just lost.
But then again, you've seen how desperate they have become advertising Core2Duo systems on their website.
Absolutely pathetic.
The inventory build is a result of a mis-match between the products they built and the products their customers wanted.
How did you misread this? What you just said is totally opposite of what was actually in the report.
Bob Rivet
in preparation for a strong fourth quarter, we built inventory by $60 million.
Dirk Meyer
Actually, most of the inventory build that we did was WIP that is being built in preparation for Q4 sales.
Hector Ruiz
we don't see any of the WIP that Dirk referred to as being anything other than strategically positioning the product for what we believe will be a strong seasonal growth in the fourth quarter.
For example, by the time Intel's chipsets will be on 90nm, most of that machinery will have been fully depreciated having been used to manufacture CPUs first.
I think you are wrong here. Name the FAB.
Yeah, they can definitely count on ATI's money losing business now that they've lost 70% of their revenue coming from intel based systems.
Well, this isn't just wrong; it's really wrong. First of all, ATI only gets half of its money from computer chips. The other half comes from consumer electronics that have nothing to do with Intel.
AMD cannot supplement what they just lost.
Absolutely pathetic.
What is truly pathetic is your understanding of the market. I doubt you'll understand this but I can try anyway.
Intel is currently unchallenged both in desktop integrated graphics and mobile integrated graphics. Intel has the lion's share of both. The strategic position of ATI is to now supply these very same integrated graphics chipsets and allow AMD to attack a market where Intel currently has no competition. I can guarantee that Intel will not be able to prevent share loss in this area. However, Intel will temper this with gains in server and desktop share.
ATI now has a guaranteed income for chipsets that it never had before. This makes ATI much more stable; there is no negative to this. With this core stability ATI is then free to pursue contracts with other OEM's and motherboard makers. It may indeed see some initial loss in revenue. However, a core production revenue that can be counted on is much better for a company of this size. It is better in terms of scheduling and inventory; it is better in terms of planning; and it is even better in terms of sales.
You also completely miss the difference in flexibility between Intel and AMD. With increased competition Intel will push their own chipsets while AMD will do either. This will tend make customers a bit annoyed with Intel.
You missed this:
Glen Yeung - Citigroup
You saw a lot of inventory build in the quarter, and I recognize at the beginning of the quarter you said you wanted to do that. But I wonder if the inventory build was exactly what you wanted, particularly given that you referenced you thought you could have done things better during the quarter?
Dirk Meyer
Actually, most of the inventory build that we did was WIP that is being built in preparation for Q4 sales. The issue within the quarter wasn't so much one of building inventory of finished goods that we couldn't sell, but rather having the right product and the right mix available for the right customer at the right time.
My comment remains. As they were trying to explain the inventory, Dirk, admitted they had a mis-match between what their customers needed and what they had in stock. Read the call transcript further please and Hector talks about how they are trying to rectify the problem. Wall Street in their earnings assesments post the earnings call and their own market checks have told their investors that AMD didn't execute as well as they have in the past which resulted in the inventory build up and this was primarily due to them not hitting the targets they had with Dell. Now you can pound Wall Street and not believe them...that's your call.
However, answer me why their channel would be short of supply while their inventory is building. The answer is two-fold:
1. They have to commit to provide Dell a certain volume. To do that, they need to hold product for Dell for the required time. Capacity implication.
2. They built the wrong products which is contributing to the inventory build up.
Well, I'm sorry but you are still misreading the transcript. There is absolutely nothing in the transcript to support the notion that AMD's current inventory is mostly made up of chips that it couldn't sell last quarter. AMD has already taken the hit for the mismatch in a 2% drop in margin. AMD has said that this 2% will be gone again next quarter. Common sense should tell you that the 2% would not disappear if AMD's inventory for next quarter included mostly mismatched chips.
Scientia - read the transcript again. Dirk's comment and Hector's explanation on how they are trying to align supply with customer demand are common sense. I'm afraid you either do not understand the real implication of their comment or are chosing to ignore it. AMD lost 5.5 points of margin and 2 points of that was by not having the product the customer wanted at the time they wanted it.
And nobody said "mostly" so I'm not sure where you're picking up that impression.
Ashenman - you need to comprehend this too. What does it mean when you say you had the wrong product at the time the customer wanted something else? It means you had product but it wasn't what the customer wanted so they didn't buy it and it got added to inventory. As a result they lost 200 BPP of margin. Hector then says we will ensure this doesn't happen next qtr and regain those 200 BPP.
Neither of you have answered my question. If the channel is short of product then why didn't AMD supply them instead of losing 200BPP of margin while building inventory that they knew would surprise Wall Street. I've already explained this so I'm not about to do it again.
Bottomline - AMD lost 5.5 points of margin QoQ. Inventory grew 15%. Wall Street has advised their clients AMD did not execute properly. AMD stock take 12% beating. AMD gains 1% share of margin. Intel gains 3%. Momentum is returing to Intel. I repeat - Momentum is returning to Intel.
If you see it differently, call Q4 as I'm about to and I'll post it here. When the results are out, we'll see who's closer.
Yup - so just to recap my position. The 60 mil inventory build is not the real issue. The issue is that it wasn't even something they could move into distribution so it must have been product built for their OEM customers (I'm guessing Dell) and one or both of the following reasons could have caused them to not move it out:
1. They have to hold inventory for Dell. Usually cancellation windows are short - maybe just a few days before the order ships. So you have to hold the Dell inventory on one hand and then the risk is if they change their build plans last minute then AMD (same issue for Intel too) get shafted.
2. Sometimes you build a specific product to the customer's requirements. Like Intel is currently doing for Apple in terms of packaging, etc to enable Apple to make even thinner/smaller clients. Which means you cannot always send that product into distribution.
AMD's growth in the last 2 years has come largely from the DIY channel through distribution. What this incident signals to me is that when you get a customer like Dell and they suddenly become such a significant part of your business, your risks and rewards become very concentrated and hence the relative volatility in your business plans goes up. It's the philosophy of not placing all your eggs in one basket.
Ashenman - in the old days if Dell and/or HP sneezed, Intel caught a cold. Because of the proportion of Intel's business with these 2 customers. Then it just became Dell because of the pace at which they grew (helped to some extent by Intel because HP had moved to AMD to dislodge Intel as a sole supplier).
Now AMD will have to watch out for that as they become POPORTIONATELY more dependant on Dell. It's not that their business collapses in one quarter if Dell changes plans...but they can miss their quarter based on Dell's decisions and performance which impacts the stock.
I decided to write an article on the Q3 earnings, "Intel's Bluff". If you are really holding onto the notion that Intel is gaining on AMD then you shouldn't read it. Although Intel's revenues should be higher than they were in 2004 Intel will actually have about $2.5 Billion less to spend this year. And, if we remove VIA from the comparison Intel actually lost share in the third quarter (both volume and revenue). The article is pretty detailed but you are welcome to refute it if you can.
You, Core2Dude and Sharikou all cherry pick whatever snippets and facts you think may support your beliefs. This is something I have never done on my blog. Cherry picking leads to a one sided point of view that also tends to be full of holes. Any assertion that I make has to be supported overall rather than just in one special case.
Post a Comment