Article: AMD's Mobile Strategy
By: David Kanter (dkanter.delete@this.realworldtech.com), December 21, 2011 10:55 am
Room: Moderated Discussions
pk (pk@notmyrealaddress.com) on 12/21/11 wrote:
---------------------------
>From the original poster:
>"3. Don't make an x86 + ARM chip. That's a dumb idea that no software engineer will like."
>
>David agreed, and I have yet to see anyone else disagree, though admittedly did
>not read all posts. But couldn't x86 + ARM be a potential differentiator for AMD
>in enabling wider compatibility than what competitors >offer?
Not really. First, it'd be a validation and complexity nightmare and second, I don't think it'd really help for tablets. You still need both x86/ARM drivers, tuning, etc. etc. which would make things tricky.
>More broadly on David's original topic of "AMD's Mobile >Strategy": why would AMD
>not pursue a merger?
AMD isn't in a very financially strong position, but they aren't in bad enough shape that they need to merge. I also can't think of any good candidates.
Part of the problem is that AMD's value is really related to x86, so whoever buys them needs to be able to make a go of it. IBM and TI already left that market.
Perhaps Samsung, Broadcom or Qualcomm want to try...but I see it as being outside Qualcomm's expertise. Many PCs get no benefit from 3G/4G wireless, and that's really their differentiator.
And as much as Samsung really wants to be a logic provider, I'm not sure they really want to be in direct competition with Intel. Remember, most of their money comes from DRAM, which requires working with Intel. Also, they might be tempted to use internal manufacturing and frankly I don't know how good Samsung's process technology is.
Broadcom seems wedded to MIPS in a big way, but who knows? I don't know a huge amount about the company.
>If they go it alone and tackle the low-power angle with >tablets
>and low-end PCs, as David suggests, they will be in a tough spot with Intel coming
>at them from above and the entire ARM ecosystem coming at them from below. They
>have reportedly killed off some 2012 SoCs and cleaned house extensively. Seems to
>fit that they would be looking at "strategic alternatives", maybe with one of the
>big wireless companies--ARM+x86+wireless+Radeon would >present an impressive breadth
>of technology and design expertise.
It would be an impressive breadth, but I think perhaps the issue is that AMD needs both breadth and depth to compete with Intel.
David
---------------------------
>From the original poster:
>"3. Don't make an x86 + ARM chip. That's a dumb idea that no software engineer will like."
>
>David agreed, and I have yet to see anyone else disagree, though admittedly did
>not read all posts. But couldn't x86 + ARM be a potential differentiator for AMD
>in enabling wider compatibility than what competitors >offer?
Not really. First, it'd be a validation and complexity nightmare and second, I don't think it'd really help for tablets. You still need both x86/ARM drivers, tuning, etc. etc. which would make things tricky.
>More broadly on David's original topic of "AMD's Mobile >Strategy": why would AMD
>not pursue a merger?
AMD isn't in a very financially strong position, but they aren't in bad enough shape that they need to merge. I also can't think of any good candidates.
Part of the problem is that AMD's value is really related to x86, so whoever buys them needs to be able to make a go of it. IBM and TI already left that market.
Perhaps Samsung, Broadcom or Qualcomm want to try...but I see it as being outside Qualcomm's expertise. Many PCs get no benefit from 3G/4G wireless, and that's really their differentiator.
And as much as Samsung really wants to be a logic provider, I'm not sure they really want to be in direct competition with Intel. Remember, most of their money comes from DRAM, which requires working with Intel. Also, they might be tempted to use internal manufacturing and frankly I don't know how good Samsung's process technology is.
Broadcom seems wedded to MIPS in a big way, but who knows? I don't know a huge amount about the company.
>If they go it alone and tackle the low-power angle with >tablets
>and low-end PCs, as David suggests, they will be in a tough spot with Intel coming
>at them from above and the entire ARM ecosystem coming at them from below. They
>have reportedly killed off some 2012 SoCs and cleaned house extensively. Seems to
>fit that they would be looking at "strategic alternatives", maybe with one of the
>big wireless companies--ARM+x86+wireless+Radeon would >present an impressive breadth
>of technology and design expertise.
It would be an impressive breadth, but I think perhaps the issue is that AMD needs both breadth and depth to compete with Intel.
David