AMD Q2 2013 Market Share up 2.2% Q to Q (Mercury Research)

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

mrmt

Diamond Member
Aug 18, 2012
3,974
0
76
Looking through the recent threads here...

...can't find Kabini in laptops.
...can't find Temash in tablets.

Yet now it was a "smooth run" in Q2?

AMD accrue revenue not when a laptop reaches the consumer or even the store shelf, but when a chip leaves AMD inventory to ODM/OEM inventory. There will be a lead time between AMD accrue revenue and you be able to buy something at the shelf. The fact that they reported higher sales is a good indicative that you should see a nice mix of Jaguar laptops in the stores soon, or at least emerging markets will.

It was a smooth run because reheated Trinity faced the same old IVB of last year and Jaguar doesn't really have competition in some market brackets. This benign scenario will start to change in Q3 and Q4, as AMD will face a much tougher competitive environment.

Haswell launched one week after Kabini. Why is Q3 going to show a Haswell impact, did the extra week really make that much difference?

What Intel launches at the top of their product line doesn't matter for AMD because they have no product to compete there. It is what Intel launches at the bottom of the chain, i3, Celeron, Pentium, that hurts AMD sales.

You should have noticed by now that when Intel launches a given chip it takes far less time than AMD's chips to appear on the market. The reason for that is that Intel ships products well before launch, while AMD should launch closer to first shipment time. And if Intel is planning a September launch for i3 Haswell, it is getting on the supply chain now, meaning that the effects on AMD sales wills start to appear in Q3.


How will Bay Trail be showing an impact in Q4, when you can't even find Temash in a tablet? Is Bay Trail now supposed to harm AMD's non-existent tablet share? None of your points make sense.

Same as before. You are thinking in terms of OEM finished products when you should think on supply chain terms. If Intel is going to launch bay trail on early Q4, impacts will start on Q3 and will be fully clear by the end of Q4.
 
Last edited:

mrmt

Diamond Member
Aug 18, 2012
3,974
0
76
My hunch is that they will lose marketshare in 2013 compared to 2012.
And the same for 2014 vs 2013.

That's what I'm expecting too. There isn't much for Richland on the mobile market and whatever server market share they still hold will be wiped out once IVB-EP arrives. Big core will be left with whatever share they can hold on the desktop market. The big question is how Jaguar will hold on the mobile market. I don't know how much pressure Intel will put on Celeron/Pentium price, but I don't think Intel can put significant pressure on that market bracket (craptops) with 22nm chips, but they will with 14nm.
 

shady28

Platinum Member
Apr 11, 2004
2,520
397
126
Looking through the recent threads here...

...can't find Kabini in laptops.
...can't find Temash in tablets.

Yet now it was a "smooth run" in Q2? Haswell launched one week after Kabini. Why is Q3 going to show a Haswell impact, did the extra week really make that much difference?

How will Bay Trail be showing an impact in Q4, when you can't even find Temash in a tablet? Is Bay Trail now supposed to harm AMD's non-existent tablet share? None of your points make sense.

Kabini is in the new Samsung Ativ Book Series 9 Lite (a lower cost series 9 Samsung Laptop).

This is also a full quarter earlier than expected, which is good for AMD :

http://www.eetimes.com/author.asp?section_id=36&doc_id=1318925
 

Lonbjerg

Diamond Member
Dec 6, 2009
4,419
0
0
That's what I'm expecting too. There isn't much for Richland on the mobile market and whatever server market share they still hold will be wiped out once IVB-EP arrives. Big core will be left with whatever share they can hold on the desktop market. The big question is how Jaguar will hold on the mobile market. I don't know how much pressure Intel will put on Celeron/Pentium price, but I don't think Intel can put significant pressure on that market bracket (craptops) with 22nm chips, but they will with 14nm.

I got the same impression looking at AMD's roadmap....it really says it all...Intel is getting to rule the high margin areas....just like they want too.
 

SiliconWars

Platinum Member
Dec 29, 2012
2,346
0
0
Bookmarked for later lulz.

You didn't need to, I'll be bringing it up believe me. There are things going on in Q3 that will be quite surprising for the uninitiated.

What Intel launches at the top of their product line doesn't matter for AMD because they have no product to compete there. It is what Intel launches at the bottom of the chain, i3, Celeron, Pentium, that hurts AMD sales.

You should have noticed by now that when Intel launches a given chip it takes far less time than AMD's chips to appear on the market. The reason for that is that Intel ships products well before launch, while AMD should launch closer to first shipment time. And if Intel is planning a September launch for i3 Haswell, it is getting on the supply chain now, meaning that the effects on AMD sales wills start to appear in Q3.
The OEM's aren't exactly falling over themselves to buy the i5 and i7 Haswells so why should this be any different? The difference between the Ivy and Haswell i3's and Pentiums isn't big enough that the Kabini's and Richlands are suddenly going to become unbuyable.

Same as before. You are thinking in terms of OEM finished products when you should think on supply chain terms. If Intel is going to launch bay trail on early Q4, impacts will start on Q3 and will be fully clear by the end of Q4.
AMD has practically zero tablet market share so Bay Trail-T isn't going to make any difference to them this year.
 
Aug 11, 2008
10,451
642
126
You didn't need to, I'll be bringing it up believe me. There are things going on in Q3 that will be quite surprising for the uninitiated.

The OEM's aren't exactly falling over themselves to buy the i5 and i7 Haswells so why should this be any different? The difference between the Ivy and Haswell i3's and Pentiums isn't big enough that the Kabini's and Richlands are suddenly going to become unbuyable.

AMD has practically zero tablet market share so Bay Trail-T isn't going to make any difference to them this year.

Rethink your math of that last sentence. If all else stays equal, and processor X sells a given amount for company A, the total market would increase, as would the market share of company A.

And actually, it seems to me that tablets is where AMD really needs to make inroads with Kabini/Temesh. We dont need any more marginally adequate laptops with Kabini and Atom (new or old) processors. They are both underpowered, and so far Kabini laptops have had very mediocre battery life for the low performance. The new Atom's will still be underpowered, but I at least expect they will have very good battery life.
 
Last edited:

AtenRa

Lifer
Feb 2, 2009
14,003
3,362
136
What Intel launches at the top of their product line doesn't matter for AMD because they have no product to compete there.

That was not what you were saying before the Haswell (core i5/i7) release. :rolleyes:

It is what Intel launches at the bottom of the chain, i3, Celeron, Pentium, that hurts AMD sales.

Yeap, but Haswell Core i3 is shipping to ODM/OEMs in Q2 for a Q3 release. That means AMD gained Market share of 1.7% in Desktop in the time of Core i3 Haswell Shipping. Also, Haswell Core i3 will only have 3 months before AMD Kaveri release in January 2014.

So again, your dooms day predictions didn't materialize. Haswell didn't impact AMDs APUs at all.

Also, Baytrail for Dektop will start shipping to ODM/OEMs after Feb 14 2014. Jaguar shipments started in Q2 2013, by the time Batrail-D starts to be released, a new Jaguar replacement will be shipping.

Intel_Atom_ValleyView_Schedule.png
 

mrmt

Diamond Member
Aug 18, 2012
3,974
0
76
You didn't need to, I'll be bringing it up believe me. There are things going on in Q3 that will be quite surprising for the uninitiated.

You mean the console chips? Whatever sales they generate it won't make too much cash for AMD because of the low gross margins.

The OEM's aren't exactly falling over themselves to buy the i5 and i7 Haswells so why should this be any different? The difference between the Ivy and Haswell i3's and Pentiums isn't big enough that the Kabini's and Richlands are suddenly going to become unbuyable.

Why do you think OEMs aren't? There is no point in buying IVB over HSW as Intel sells the latter for the same price. So OEMs will buy i7 and i5 for their premium line up.

As for HSW i3 vs Kabini vs Richland, I'm not saying that HSW will make AMD chips "unbuyable", but that it will put pressure on AMD pricing. Haswell power saving features puts it on another category on the mobile market and it is something that AMD cannot match at the same performance level. It won't outclass Kabini because this one has lower performance but has a nice cost advantage, but it will outclass Richland for everyone but the budget gamer, because this one has similar performance but none of the features. The only advantage will be a much eroded GPU performance advantage.

And Haswell is a Tock, and Tocks are meant to bring the latest technological package to the market. Broadwell is a Tick, and a Tick brings the latest technological package to the market within a better cost structure. Just look at the impact (nehalen vs westmere, sandy bridge vs ivy bridge) on AMD sales. Ticks always have a bigger impact for AMD because of the better cost structure.

AMD has practically zero tablet market share so Bay Trail-T isn't going to make any difference to them this year.

I expect Intel to put Bay Trail on craptops, and while Kabini is the right chip for craptops, it will create some kind of pressure on the market bracket.
 

mrmt

Diamond Member
Aug 18, 2012
3,974
0
76
The difference between Kabini and the new Atoms will be basically nothing. Whatever slight single thread performance lead the Atom has will be lost in graphics. The Atom is 10W TDP and we know the A4-5000 Kabini is basically sub-10W, so battery life will not be much different either.

And here you just answered why AMD will have a harder Q4. Right now Kabini doesn't have anything close to the kind of competition Bay Trail will provide. And while it will be the right chip for craptops, no competition always yields better margins than any competition.
 

lefty2

Senior member
May 15, 2013
240
9
81
How did AMD manage to gain markeshare in servers then? As far as I know, they had no new server products to that quarter.
 

NTMBK

Lifer
Nov 14, 2011
10,486
5,908
136
How did AMD manage to gain markeshare in servers then? As far as I know, they had no new server products to that quarter.

Did they not launch the Opteron X chips in that quarter? (The rebadged Kabini parts.)
 

mrmt

Diamond Member
Aug 18, 2012
3,974
0
76
How did AMD manage to gain markeshare in servers then? As far as I know, they had no new server products to that quarter.

Once you go to such low volumes as AMD server business QoQ comparisons become meaningless because one big order has the potential to distort the entire statistical trend. 50 million in servers for Intel is 2% of the >>server business<<, 50million in servers for AMD is almost 5% of the >>entire company<<. So if you want to check the health of AMD server business, you must check the numbers in three or four quarters.

Btw, Kabini is also a server product, so yes, they did have a new server product on sale.
 

Slomo4shO

Senior member
Nov 17, 2008
586
0
71
Why isn't the thread title : AMD Q2 2013 Market Share up 2.2% Q to Q but down 1.1% compared to 2012?

Is it about as relevant as the 9.5% increase in total chip sales Q to Q and the 8.5% decline compared to 2012.
 

erunion

Senior member
Jan 20, 2013
765
0
0
The difference between Kabini and the new Atoms will be basically nothing. Whatever slight single thread performance lead the Atom has will be lost in graphics. The Atom is 10W TDP and we know the A4-5000 Kabini is basically sub-10W, so battery life will not be much different either.

You continually equate TDP to battery life. Stop it, just stop it. TDP is not representative of average power draw.

Kabini lacks the low power idle states that clovertrail and baytrail have.
 

SiliconWars

Platinum Member
Dec 29, 2012
2,346
0
0
You continually equate TDP to battery life. Stop it, just stop it. TDP is not representative of average power draw.

Kabini lacks the low power idle states that clovertrail and baytrail have.

So why no "SDP" on the M parts?

Battery life while it sits doing nothing is meaningless, and the inherent lack of performance in these chips means they spend most of their time at load even when doing something simple like browsing.
 

PPB

Golden Member
Jul 5, 2013
1,118
168
106
Idle battery life numbers are for the usual PR talk. I'm more concerned of battery life in these situations:

1° Gaming: Laptop shutting down after only 1 match of LoL is a no-no for me.
2° Video playback/youtube.
3° Multitasking with low GPU load.

Dont know about you guys, but I actually want to use my laptop for long periods of time, not just watch it idle and think "omfg look at dat battery laif!!!1!one!eleventy".

And it has been shown that at least for pre-Haswell architectures (still didn't see the updated test with the new platform) from Intel, things get nasty when you are in the scenario 1 I described.
 
Aug 11, 2008
10,451
642
126
Idle battery life numbers are for the usual PR talk. I'm more concerned of battery life in these situations:

1° Gaming: Laptop shutting down after only 1 match of LoL is a no-no for me.
2° Video playback/youtube.
3° Multitasking with low GPU load.

Dont know about you guys, but I actually want to use my laptop for long periods of time, not just watch it idle and think "omfg look at dat battery laif!!!1!one!eleventy".

And it has been shown that at least for pre-Haswell architectures (still didn't see the updated test with the new platform) from Intel, things get nasty when you are in the scenario 1 I described.

I was talking about low power/performance laptops such as Kabini/Atom, 11.6 in types. Honestly, I dont care about gaming battery life on that type of machine. If I want to game on a laptop I would use a more powerful one and just plug it in. Scenarios 2 and 3 perhaps, but again I cant see using one of these for much multi-tasking either. My main interest in a laptop like this is for e-mail and web browsing on the go, with a bit more flexibility than an ARM device and enough idle/sleep battery life to not have to turn it on and off too frequently. I saw a kabini laptop advertised in the Best Buy ad this week, but it had only 5 hours advertised battery life, which for such a low performance device is basically a no go for me. I mean, even my work laptop with a full power i5 has almost that much battery life.
 

pablo87

Senior member
Nov 5, 2012
374
0
0
As others have said, the Server #"s are meaningless.

Also, Kabini and Temash should not be grouped together. Temash while a good product, just won't cut it in the tablet space. Neither will BT but Intel can at least throw $$$ at the problem.
 

krumme

Diamond Member
Oct 9, 2009
5,956
1,596
136
As others have said, the Server #"s are meaningless.

Also, Kabini and Temash should not be grouped together. Temash while a good product, just won't cut it in the tablet space. Neither will BT but Intel can at least throw $$$ at the problem.

Agree, Intels market share in tablets will therefore reflect how many $$$ they throw at it. The same aplies for the phones. I expect them to do that, as they have already invested a lot a money and prestige here. Its a dead end and the hard reason to be there is still beyond my understanding.

That leaves Intel and AMD fighting for the craptop market now and in 1-4 years perspective. Kabini & Temash, and successors, is as said probably a better fit here, but Intel have the $$$. Margins is, at best, going to be razor thin for AMD.

Bringing ARM to the table, here as well as for servers in the 4-8 years perspective, looks like the only long term winning strategy for AMD , because it removes some of Intels monopoly advantages. But they have to do that without $$$*, and in a mature market where the incentive to pay for cpu/gpu power becomes less and less for each year.

Isnt that what you call a challenge ? lol

*maneuvering a company without financial headroom is a nightmare imho, and i have the deepest respect for those who do it.
 

mrmt

Diamond Member
Aug 18, 2012
3,974
0
76
Bringing ARM to the table, here as well as for servers in the 4-8 years perspective, looks like the only long term winning strategy for AMD , because it removes some of Intels monopoly advantages. But they have to do that without $$$*, and in a mature market where the incentive to pay for cpu/gpu power becomes less and less for each year.

You forget that sometimes Intel monopoly advantages work in AMD's favor.

There was always places in the x86 market that weren't interesting enough for Intel and these places were left for AMD, Cyrix and the other little ones that had x86 licenses, and nobody outside the club could go for that market and the club is basically AMD today. AMD lived in Intel shadows for a long time selling Intel clones on n-1 node and they have more or less reversed to this strategy on every other segment but Jaguar, where their products offer reasonable trade offs when compared to Intel's.

With ARM AMD loses this protection. There will be always some shady designer offering "me too" solutions at n-1 or n-2 nodes, and I cannot fathom easier competition from Samsung and Qualcomm than from Intel. AMD basically doesn't get rid of the 800 pound gorilla in their room but they also lose whatever protection they enjoyed from Intel IP umbrella.

Maybe being 2 nodes behind would make most of AMD line up nonviable, maybe they think they can fight Qualcomm and Samsung better than they can fight Intel, or maybe ARM is just a better gateway for the kind of embedded business they want to build for themselves... the fact is that there are risks associated with this decision and they are not small.

My guess is that as Vanilla ARM provides a much cheaper alternative than Jaguar for that. And despite the hype in microservers and others, it is embedded that will set the direction for the company, as it will answer for 50% of their business in the medium term.
 

SiliconWars

Platinum Member
Dec 29, 2012
2,346
0
0
You seem to have "forgotten" that AMD and Qualcomm/Samsung are using very similar nodes at their competition point. Samsung's is behind in fact.

AMD is clearly heading towards ARM cores there anyway, which means they'll be the only company able to offer ARM and x86 if they so choose.
 

mrmt

Diamond Member
Aug 18, 2012
3,974
0
76
You seem to have "forgotten" that AMD and Qualcomm/Samsung are using very similar nodes at their competition point. Samsung's is behind in fact.

Hmmm... Are you suggesting that once AMD is in the same node of the lead pack of the ARM crowd whatever handicap they have in design will be less pronounced than in x86, where AMD has both a node and design handicap? If this is the case, you might be right. As they are in the same node whatever cost advantage stems more from the commercial deal and R&D structure and not because of Intel fancy nodes that offers much better cost structure as in x86.

But node aside, Samsung and Qualcomm can put much more money on chip R&D than AMD can, and that will reflect rather sooner than later, when ARM SoCs grow in size and complexity.


AMD is clearly heading towards ARM cores there anyway, which means they'll be the only company able to offer ARM and x86 if they so choose.

And since when this is an advantage? It would be an advantage if AMD could offer something competitive in both fronts. They clearly can't on x86, will a vanilla ARM core be enough on ARM?