Our Thoughts

Obviously each of the four companies we talked about today had their own take on the merger and no one here can really predict where things will go.  On the one extreme, this deal could completely change everything; the combination of AMD and ATI could result in a powerhouse finally able to really compete with Intel while at the same time taking serious market share away from NVIDIA.  On the other extreme, the combination could be a mess, resulting in a bloated company that is not much better off than before, with Intel and NVIDIA standing to reap the greatest rewards.  It's far too early to tell and with the deal not even closed yet, there's no sense in counting any chickens. 

That being said, there are some points we feel are worth discussing, and some analysis that can help give perspective to the whole merger.  A lot of what we've presented so far is the company line from the four major CPU/Chipset manufacturers, and clearly there is going to be some bias in how each views the situation.  It's far too easy to look at only one side of this story, so the following is designed to help examine the pros and cons of a handful of the facets of this merger.

Integrated Graphics is about to get Better

There's one definite outcome from the AMD/ATI deal, and that is an increase in the quality of integrated graphics.  Currently, Intel is relatively unchallenged when it comes to integrated graphics solutions; independent GPU manufacturers like ATI and NVIDIA offer better integrated graphics solutions, but at much higher prices to keep their margins high.  If AMD acquires ATI, we could in theory see an AMD platform with integrated graphics that would significantly outperform a similarly priced integrated Intel platform - hopefully this will force Intel to improve its integrated graphics. 

Taking things one step further: if AMD puts a small ATI graphics core on-die or at least on-chip, the performance expectations for integrated graphics solutions would go up once more.  Remember that one strength consoles have is the extremely low latency, high bandwidth interconnect between the CPU and GPU; a highly integrated CPU/GPU combo would be able to deliver that same performance advantage on a relatively low end PC. 

Cool Stuff is Coming

On the investor call where AMD officially announced its plans to acquire ATI, a common theme discussed was AMD's Torrenza strategy.  As AMD announced at its analyst day back in June, AMD plans on openly licensing its coherent HyperTransport bus allowing for companies like ATI or NVIDIA to develop GPUs and other co-processors that would plug into a Hyper Transport slot on a motherboard.  The benefit is a very low latency, cache coherent interface between the CPU and any other device that it needs to feed large amounts of data to.  With ATI operating under AMD's wing, AMD effectively guarantees that we'll see GPUs take advantage of Torrenza. 

Intel’s Position: The Silent Treatment Our Thoughts: Will AMD manufacture ATI GPUs?
Comments Locked

61 Comments

View All Comments

  • Zebo - Wednesday, August 2, 2006 - link

    No worries...AMD runs a tight, efficient company that is accustomed to surviving
    through very hard times. AMD survived for a long time making chips that were cheap and almost as powerful as Intel's best. If they have to fall back to that business model to survive, they will. I personally loved those days of $40-$80 chips. But that's not realistic considering where AMD has been, their name and market presence currently, products on the table.. AMD is a mainstream player now with good reputation and large OEM's building thier boxes with them. They aint going anywhere.
  • poohbear - Wednesday, August 2, 2006 - link

    so if the deal goes through, will the ATI brand name disappear? will we see AMD graphics cards instead of ATI graphics cards?
  • Sunrise089 - Wednesday, August 2, 2006 - link

    IMHO there would be no reason to abandon the second most valuable GPU name. When Ford bought Aston Marton they didn't suddenly rename the products things like Ford DB7.
  • erwos - Wednesday, August 2, 2006 - link

    Let me toss out a few random thoughts. I'm more of an economist than a businessman, but I took enough banking and finance to know enough to hurt myself.

    Almost all huge corporate mergers are not huge successes. Indeed, most of them tend to be failures unless the businesses are _very_ similar (gold mining company A buys out gold mining company B). My favorite example is Novell buying out SuSE and Ximian - everyone's doing operating systems, yet the best you can say was that it wasn't a complete failure. Certainly, the promised benefits haven't really emerged. Another good example is AOL and Time Warner.

    The bad news here is that ATI and AMD are in two different sections of the industry, and that for the proposed benefits of this merger to work, they're going to have to integrate very tightly. To make things worse, the benefits of integration aren't all that clear. GPU on a CPU? Who's been asking for that? It has certain implications for the embedded market (think Geode and system on a chip applications), but they hardly needed to buy a company the size of ATI to accomplish that particular goal. And it couldn't be to hand ATI the better fabs, either - as Anand pointed out, AMD isn't going to have any extra fab space in the medium-term outlook.

    My prediction: ATI-AMD will spend the next 9 months after the merger at _vastly_ decreased efficiency. Intel and nVidia will both be able to exploit this and take definitive leads in technology, at least for a while. In the long-term, ATI-AMD's dedication to high-end GPUs will fade, because the former-AMD executives running the company have absolutely no experience in the field. I am pessimistic, because, unfortunately, that is the historical truth.

    Personally, I think that if GPU on a CPU becomes the prevailing way to go, nVidia will just buy out VIA or Transmeta. And they'll probably have just the same problems as ATI and AMD will have, too... But there's no reason to toss those problems on yourself until you have to, and there was no really compelling reason for AMD to buy ATI at this moment in time.
  • Kim Leo - Wednesday, August 2, 2006 - link

    what are you talking about? ok its fine to comparte other situation like this, but AMD didnt buy ATI just for the "intergrated graphics in CPU" idea and even though Hector Ruiz dosn't have too much experience in this sector but ATI's CEO who will still be there does, and i don't think that AMD won't listen to what he has to say about it. I think this will be great, AMD and ATI will both benefit from this, they both get technologies that can be used in their own products
  • erwos - Wednesday, August 2, 2006 - link

    There aren't two CEOs. There's one, and his name is Hector Ruiz. At best, ATI's CEO will get pushed into director of the graphics division. More to the point, AMD's the much bigger company, and it's more likely their corporate culture is going to dominate ATI's. ATI's CEO's opinion will matter, but it's not going to sway AMD like it did/does ATI.

    If the plan isn't to integrate GPUs on CPUs, what other benefit was there to acquiring ATI? What techologies is ATI going to give AMD, and vica versa?

    -Erwos
  • Sunrise089 - Wednesday, August 2, 2006 - link

    Couldn't the desire to purchase a healthy company with a high profit margin in a fast growing industry be a benefit? I think everyone is too focused on integration in the short term. AMD had $$$, $$$ is there to spend or invest, and if the bean-counters at AMD think the ROI for buying ATI is higher than investing in a new fab or whatever than they make that decision.
  • JarredWalton - Wednesday, August 2, 2006 - link

    The major benefit seems to be AMD getting a company with a reasonable chipset business, and they can work that to create better business platforms, thus helping to penetrate the lucrative business sector. Except, penetrating the business sector is extremely difficult, especially the corporate world. "Buy Intel and Dell" is the standard decision, and even if Dell isn't picked, almost all businesses buy Intel systems. They did this all through the "NetBurst failure", so why would they change now that Intel has a good chip again (Core 2)?
  • yacoub - Wednesday, August 2, 2006 - link

    So will we see a reference cooler design on future ATI cards that is less noisy than the silly thing on the X1800/X1900 series? ;P
  • jones377 - Wednesday, August 2, 2006 - link

    In Q106 the marketshare breakdown for all x86 chipsets were as follows....

    Intel 57%
    VIA 15%
    ATI 12%
    Nvidia 9%
    SiS 6%

    http://www.xbitlabs.com/news/chipsets/display/2006...">http://www.xbitlabs.com/news/chipsets/display/2006...

    Different breakdown for Intel and AMD platforms. Basically Nvidia has almost no share in the Intel platform market while ATI sells in both.

Log in

Don't have an account? Sign up now