Richland & Kabini rumours

[SiliconWars]

ASP is more likely to be in the $40-$80 range. That these will be A4 and A6 branded tells it's own story.

If the entire SoC is 100mm2, that's 600 die per 300mm wafer. At 80% yield that would be 480 good die ~$60 = $28800 per wafer at a cost of $6k per wafer.

You can think what you want but it's clearly going to put AMD in much better shape than they currently are selling salvaged 240mm2 dies for their current A4's and A6's.

 

[mrmt]

ASP is more likely to be in the $40-$80 range. That these will be A4 and A6 branded tells it's own story.

Mercury Report announced 16 MM processors for AMD in Q312, they reported 927 million in sales in the CPU division. Doing simple math, you'll get an ASP around $57. The CPU division also includes Seamicro and Opteron processors, both of which sell SKUs far above $57. So it's safe to say that ASP for the consumer business are below $55, even below $50.

In the consumer business, Llano, Bulldozer and Trinity still command higher prices than Brazos (regardless of their gross margins). Once you factor all this I don't think you may get a $80 ASP by any means. ASPs around $25-$35 for Brazos are far easier to fathom.

If the entire SoC is 100mm2, that's 600 die per 300mm wafer. At 80% yield that would be 480 good die ~$60 = $28800 per wafer at a cost of $6k per wafer.You can think what you want but it's clearly going to put AMD in much better shape than they currently are selling salvaged 240mm2 dies for their current A4's and A6's.

Don't get me wrong, regardless of performance comparisons, I do think that Kabini will have *much* better economics than Trinity or Bulldozer, not much for Kabini merits but because Bulldozer was a train wreck in the form of silicon. What I'm pointing out is that, for the PC market, Kabini does not solve the structural problems AMD has. If AMD can put Kabini on embedded systems, then it will be a home run, but Kabini on the PC is a temporary relief. Intel is keen to refresh both their product lines in a pace AMD cannot match. It will be a matter of time until AMD is hopelessly behind with Kabini too.

 

[Nemesis 1]

To break even they only need to make $3,300,000,000. Kabini and Temash will be very successful in comparison to VIA's new 28-nm monolithic quad-core and Intel's 32-nm/22-nm Atoms.

Ya see this just isn't right . The 32nm process being listed from intel I have no problems with . But the 22nm process you and I and everyone else is clueless on what kind of performance Silvermont will bring in 2013 . It is being released in 2013. Were already in 2013 . Next weeks show should actually give all of us some talking points . But hyping 1 unreleased product over another unreleased Product is shameful and yes I do it as well . The problem here is The AMD people are picking AMD over intel in the processor market . The likely hood of that is SMALL . It may very well be faster than silvermont I haven't seen a comparison . But here is the Problem it will nor compete with intels 10 watt. haswell. Only on price it will not compete with arm devices in efficiency and certainly not Silvermont as this will be the best in efficiency. It amazes me that we know more about haswell . Yet with haswell we don't not know how many EU will be used . Yet with silvermont someone some where mentioned a 4 eu chip . The only way I except that is if the phone model gets the 4EU part . As for AMD beating intel to market . It will be vary close .
Here a copy from AT article on Intel atom (saltwell) . Sillvermont is a 2013 launch product . The Phone will be first followed by tabs . Intel could release the Server Silvermont first . But I think it will be the same order as last year .

Here is that copy paste from AT 12/12/12.

Clearly Silvermont is a 2013 product that many seem to be dening here.

In 2013 Intel will release Avoton, Centerton’s successor and based on Intel’s 22nm process. Based on Intel’s previous roadmaps we know that 22nm is also supposed to coincide with the launch of Intel’s new Silvermont architecture, so it’s reasonable to assume that Avoton will be Intel’s Silvermont-based processor for servers. Processors based on the ARMv8 ISA are not expected to launch until late in the year, so it’s possible that ARM and its partners will be going up against Avoton rather than Centerton/S1200.

 

[Nemesis 1]

Here is more about Silvermont . There is also some benchies so its in the wild . So next week promises excitement for intel fanbois like me .

http://www.intomobile.com/2012/05/14/intel-merrifield-system-chip-may-end-up-2013-superphones/

 

[SiliconWars]

Mercury Report announced 16 MM processors for AMD in Q312, they reported 927 million in sales in the CPU division. Doing simple math, you'll get an ASP around $57. The CPU division also includes Seamicro and Opteron processors, both of which sell SKUs far above $57. So it's safe to say that ASP for the consumer business are below $55, even below $50.

In the consumer business, Llano, Bulldozer and Trinity still command higher prices than Brazos (regardless of their gross margins). Once you factor all this I don't think you may get a $80 ASP by any means. ASPs around $25-$35 for Brazos are far easier to fathom.

4 core Llano ASP was $80, Trinity will be similar. 2 core Llano was $60. There's no reason why Kabini won't be selling for $40-$80.

Don't get me wrong, regardless of performance comparisons, I do think that Kabini will have *much* better economics than Trinity or Bulldozer, not much for Kabini merits but because Bulldozer was a train wreck in the form of silicon. What I'm pointing out is that, for the PC market, Kabini does not solve the structural problems AMD has. If AMD can put Kabini on embedded systems, then it will be a home run, but Kabini on the PC is a temporary relief. Intel is keen to refresh both their product lines in a pace AMD cannot match. It will be a matter of time until AMD is hopelessly behind with Kabini too.

Intel has no competition now or in the near future. Atom will still be a joke - I guarantee you that when the 22nm Atom is compared to Kabini it will look even worse than the 45nm did vs Brazos. While it will have superior TDP characteristics, it will be because of having pathetic graphics. There is zero reason to believe anything will change from the current Atom vs Brazos situation.

 

[mrmt]

4 core Llano ASP was $80, Trinity will be similar. 2 core Llano was $60. There's no reason why Kabini won't be selling for $40-$80.

Disclose your source, please.

 

[SiliconWars]

Disclose your source, please.

http://investorvillage.com/mbthread.asp?mb=476&tid=10759150&showall=1

 

[Nemesis 1]

4 core Llano ASP was $80, Trinity will be similar. 2 core Llano was $60. There's no reason why Kabini won't be selling for $40-$80.

Intel has no competition now or in the near future. Atom will still be a joke - I guarantee you that when the 22nm Atom is compared to Kabini it will look even worse than the 45nm did vs Brazos. While it will have superior TDP characteristics, it will be because of having pathetic graphics. There is zero reason to believe anything will change from the current Atom vs Brazos situation.

Are we talking about 17 watt processors if so . They aren't going against Intel Atom but haswell . If your talking below 10 watts . lets wait and see , Atom is at most a 10 watt chip and that I believe is server chip . If the comparison is made at same wattage I think Intel wins . But ignorer to go over the 10 watt threshold Atom isn't the compare chip Haswell is . If people are after power they will buy haswell. If they are after efficiency than Atom silvermont will win . AMD can only fit between the 2 intel products . Basicly Intel has ARM on the ropes at 22nm silvermont and. and Intel has AMD between a rock and a hard place . Only a blind man couldn't see this .

 

[mrmt]

http://investorvillage.com/mbthread.asp?mb=476&tid=10759150&showall=1

This is Q211 data, when AMD had ~35% more in revenues, 25% more shipments and gross margins was a whopping 8% higher. They are really outdated. The numbers I brought here are from Q312.

 

[SiliconWars]

Are we talking about 17 watt processors if so . They aren't going against Intel Atom but haswell

They are not going against Haswell. Kabini will sell 30 million+ units per year, intel will be lucky to sell 3 million 17W Haswells per year.

This is Q211 data, when AMD had ~35% more in revenues, 25% more shipments and gross margins was a whopping 8% higher. They are really outdated. The numbers I brought here are from Q312.

Yes and the reason for all of that is that they had to practically give the remaining Llano's away after they were Osborned by Trinity.

 

[Nemesis 1]

Who needs up to date facts in a slug fest? lol

 

[mrmt]

Yes and the reason for all of that is that they had to practically give the remaining Llano's away after they were Osborned by Trinity.

According to AMD management ASP for CPUs fell across the board in the last few quarters, Llano is just the worst performer of them, as they are competing internally with Trinity and as a result are being sold below cost now, but by no means the others escaped unharmed. Bobcat sales fell on Q3 btw.

If in Q211 when things were rosier and AMD was making money, Brazos ASP was between $25-$35 (and obviously with costs well above 10$), you can bet that Kabini won't go much further on the price ladder.

 

[SiliconWars]

According to AMD management ASP for CPUs fell across the board in the last few quarters, Llano is just the worst performer of them, as they are competing internally with Trinity and as a result are being sold below cost now, but by no means the others escaped unharmed.

Yes it was a bad quarter, we get it. Would you like to bet that Q4 was better?

Bobcat sales fell on Q3 btw.

Bobcat is 2 years old and it's successor is only a few months away. How many Atoms sold?

If in Q211 when things were rosier and AMD was making money, Brazos ASP was between $25-$35 (and obviously with costs well above 10$), you can bet that Kabini won't go much further on the price ladder.

Explain how costs are "obviously" well above $10 on Brazos.

 

[mrmt]

Yes it was a bad quarter, we get it. Would you like to bet that Q4 was better?

I could bet, but that would be unfair with you. According to AMD own management they are forecasting another 10% drop in their sales, and they are going to take another GLF charge that will crush their gross margins, but I expect gross margin to shrink even if you factor out this charge. They also stated that they are going to have losses and negative cash flow. If you still want to bet, be my guest.

Bobcat is 2 years old and it's successor is only a few months away.

I don't disagree with that. I just said that the ASP you speculated for Kabini 40-80, is far out of touch with Brazos numbers when the chip was launched. In Q211, when the chip was fresh from the factories, ASP was around $25-$35. There is nothing pointing out to Kabini more than doubling this value as you implied.

Explain how costs are "obviously" well above $10 on Brazos.

Because obviously AMD gross margins didn't shoot through the roof when Brazos started selling in Q111, much less in Q212 when Brazos was the most important SKU in their mobile line up. Quite the opposite, gross margin went down slowly. And while part of the decline can be credited to Llano and Bulldozer in servers, the fact that margins didn't go significantly up means that the gross margin isn't that far from the average. If you factor the kind of market that Brazos compete and the lack of clout with OEMs, Brazos margin should be a bit lower than the average. That would put gross margin in the 40-50% range. Brazos wasn't irrelevant for the shipments there,

Given the numbers you already brought here, a combined price of around $32 at launch date and applying the assumption in the above paragraph, we get something between $16 to $19 for the average unit cost. That's much better than Llano or Bulldozer will ever be, but it's not the spectacular dream some people expect.

Your assumption, that Brazos costs around $10 and sold for around $32 would make gross margins for Brazos around 70%.

 

[Abwx]

Just came from a general store a la Walmart here in France...

There was no quad core offering set apart a i5 laptop , all the rest
including desktops was dual core using pentiums and even bobcats ,
the most refined are using i3s with radeon gfx...

It appears that pc market is rapidly being comoditized with low selling
price as the end of all and in this respect AMD s Read was right when
pointing that consumers dont mind much about the CPU as they know
that whatever they choose it will be enough for many years before
being replaced by an even cheapest one...

In this respect , kabini will fit perfectly the needs not only of western
high income users markets but also the one of emerging markets users
that are on a tighter budget.

 

[SiliconWars]

I could bet, but that would be unfair with you. According to AMD own management they are forecasting another 10% drop in their sales, and they are going to take another GLF charge that will crush their gross margins, but I expect gross margin to shrink even if you factor out this charge. They also stated that they are going to have losses and negative cash flow. If you still want to bet, be my guest.

I guess we'll just have to wait and see.

I don't disagree with that. I just said that the ASP you speculated for Kabini 40-80, is far out of touch with Brazos numbers when the chip was launched. In Q211, when the chip was fresh from the factories, ASP was around $25-$35. There is nothing pointing out to Kabini more than doubling this value as you implied.

Brazos was also given "essential" branding, not mainstream. Do you think that AMD is just going to give Kabini A6 and A4 branding, but not A6 and A4 pricing? Why would they do that?

Because obviously AMD gross margins didn't shoot through the roof when Brazos started selling in Q111, much less in Q212 when Brazos was the most important SKU in their mobile line up. Quite the opposite, gross margin went down slowly. And while part of the decline can be credited to Llano and Bulldozer in servers, the fact that margins didn't go significantly up means that the gross margin isn't that far from the average. If you factor the kind of market that Brazos compete and the lack of clout with OEMs, Brazos margin should be a bit lower than the average. That would put gross margin in the 40-50% range. Brazos wasn't irrelevant for the shipments there,

Given the numbers you already brought here, a combined price of around $32 at launch date and applying the assumption in the above paragraph, we get something between $16 to $19 for the average unit cost. That's much better than Llano or Bulldozer will ever be, but it's not the spectacular dream some people expect.

Your assumption, that Brazos costs around $10 and sold for around $32 would make gross margins for Brazos around 70%.

Actually the real reason for the slowly dropping margins over the past 2 years would be the complete cratering of server prices. The only reason margins remained fairly static for so long was due to much better margins on Brazos. It should be pretty clear to anyone that Brazos is the only chip AMD has that is worth manufacturing, and it is keeping them afloat. If Brazos margins were as low as you claim, the overall margin would have plummeted long before now.

Didn't you say yourself that Brazos sales were down (13%) last quarter? And margins down 8%? It's not exactly hard to see the correlation.

 

[Nemesis 1]

I could bet, but that would be unfair with you. According to AMD own management they are forecasting another 10% drop in their sales, and they are going to take another GLF charge that will crush their gross margins, but I expect gross margin to shrink even if you factor out this charge. They also stated that they are going to have losses and negative cash flow.

I had the exact same thoughts . It funny I think it would be great if AMD has a successful money making CPU. Its just that fitting between 2 intel processors Big core little core I just don't see it happening . The way tablets are selling AMD needs a strong tablet . Under 10 watts I not seeing it.

 

[SiliconWars]

I had the exact same thoughts . It funny I think it would be great if AMD has a successful money making CPU. Its just that fitting between 2 intel processors Big core little core I just don't see it happening . The way tablets are selling AMD needs a strong tablet . Under 10 watts I not seeing it.

You've heard the rumours that Temash is in Microsoft's Surface Pro, right?

 

[Abwx]

Under 10 watts I not seeing it.

A few days left before ULP Kabini/Temash are to be demonstrated and you are
the only one who see nothing coming , this is telling about how much
you are cluless and just trolling all the way...

Posted a few pages earlier in this thread :

http://www.hardware.fr/tag_flux/597/temash.html

 

[mrmt]

Actually the real reason for the slowly dropping margins over the past 2 years would be the complete cratering of server prices. The only reason margins remained fairly static for so long was due to much better margins on Brazos. It should be pretty clear to anyone that Brazos is the only chip AMD has that is worth manufacturing, and it is keeping them afloat. If Brazos margins were as low as you claim, the overall margin would have plummeted long before now.

Not only servers, but they are being slowly pushed down and down in both mobile and desktop price ladder. I rest my case regarding Brazos COGS.

Didn't you say yourself that Brazos sales were down (13%) last quarter? And margins down 8%? It's not exactly hard to see the correlation.

Llano was selling under cost and server is almost wiped out. Given Brazos share in AMD revenue, a mere 15%-20% of the CPU group, there is no point in this being caused by Brazos alone.

 

[Nemesis 1]

You've heard the rumours that Temash is in Microsoft's Surface Pro, right?

Yep I have and I hope they get it . Intel doesn't need MS . MS needs Intel . Intel really doesn't care what os it runs under. Apples high pricies are on their way out sooner than latter . MS overcharges for its software . Intel is being really sneaky. I hope its more than what is being said . As we know Intel prepared referance designs for phones and tablets.
It is my hope that Intel starts making its own phones intel products . Look at Apple or samsung . Intel needs to have its own brands this is where the real money is. Intel doesn't need the others. On the phones thats more than true . That doesn't stop intel from selling to Oems. It doesn't matter they wouldn't like it . The more intel puts on the die the more they need to sell a Intel branded phone. Than tablets than the worlds most power and wealthy company.

 

[ShintaiDK]

You've heard the rumours that Temash is in Microsoft's Surface Pro, right?

I sure dont hope you believe those rumours.
http://www.microsoft.com/Surface/en-US/surface-with-windows-8-pro/specifications

 

[Nemesis 1]

A few days left before ULP Kabini/Temash are to be demonstrated and you are
the only one who see nothing coming , this is telling about how much
you are cluless and just trolling all the way...

Posted a few pages earlier in this thread :

http://www.hardware.fr/tag_flux/597/temash.html

So when we going to see this mircle of AMD engineering . At CES 2013 . I won't hold my breath on that one . And your thinking is missed up . Under 10 watts Amd Temesh is nothing that Silvermont can't handle . Under 10 watts AMD has not shown a gpu that can compete with intel Hd4000 gpus . So where oh where did AMD suddenly get the engineers that are better than intels? Easy answer they don't . Your creaming youir panties over an AMD cpu that is magicly going to jumps intels OoO silvermont in the sub 10 watt arenea LOL . Braso was able to get better performance than intel simply because Intel was a sitting target. Does a braso outperform the 32nm dual core atom . I thought AT article said know it doesn't .

 

[SiliconWars]

I sure dont hope you believe those rumours.
http://www.microsoft.com/Surface/en-US/surface-with-windows-8-pro/specifications

Surface Pro 2...

So when we going to see this mircle of AMD engineering . At CES 2013 . I won't hold my breath on that one . And your thinking is missed up . Under 10 watts Amd Temesh is nothing that Silvermont can't handle . Under 10 watts AMD has not shown a gpu that can compete with intel Hd4000 gpus .

Where is the sub 10W HD4000 part for comparison?

Here is a 2 year old 9W C-50 gaming -

http://www.youtube.com/watch?v=TTf1oYM78yM
http://www.youtube.com/watch?v=TI7pLN1Bvqs

You think intel is even close at 9W? They are barely competitive at double that. Temash will be faster at <6W.

 

[ShintaiDK]

Surface Pro 2...



Where is the sub 10W HD4000 part for comparison?

Here is a 2 year old 9W c-60 gaming -

http://www.youtube.com/watch?v=TTf1oYM78yM
http://www.youtube.com/watch?v=TI7pLN1Bvqs

You think intel is even close at 9W? They are barely competitive at double that. Temash will be faster at <6W.

Temash will not be in a Surface Pro 2. It always the same rumours isnt it? Apple didnt go with the super good APUs either. It was else a done deal!