19 Comments

Summary:

Updated at the end: The way we use computers is changing, as device makers and users emphasize mobility and incredible graphics. I’ve argued that these trends signal the end of x86 computing, but what I’ve ignored is Intel’s drive to bring its brand of x86 computing […]

Updated at the end: The way we use computers is changing, as device makers and users emphasize mobility and incredible graphics. I’ve argued that these trends signal the end of x86 computing, but what I’ve ignored is Intel’s drive to bring its brand of x86 computing to these markets, which are traditionally based on other instruction sets. If it succeeds, we may see Intel inside everything from our mobile phones to our set-top boxes.

Rivals such as ARM, which is licensing its intellectual property to a variety of chip firms that build application processors for mobile phones, or Nvidia, which is helping devleopers write code to run more applications on its graphics processing chips, are showing how little the CPU matters when it comes to popular new computing paradigms such as delivering HD video or controlling battery-powered smartphones.

However,  Intel is also expanding into graphics processors and low-power computing for wireless devices and embedded systems with x86 based chips. Atom is a low-power x86 chip that Intel will use to duke it out against ARM in low-power and embedded systems, and Larrabee will be an x86 GPU that will compete with graphics chips from AMD and Nvidia.

Broadpoint AmTech analyst Doug Freedman thinks Intel can do it, especially given the chipmaker’s aggressive push into these new markets — mobile phones, embedded processors and even GPUs. In a note published today upgrading Intel, he said, “We believe long-term fundamentals of multi-year share gains in non-PC markets through Atom-based SoC solutions will enable significant (10 percent or more incremental market opportunity) new revenue opportunities.”

He also points out that Intel’s very control of the x86 chips inside computers could hinder GPU rivals once Larrabee comes out, as those GPU makers may not have the appropriate licenses to tie their chips to Intel’s CPUs. But the PC will be just one small (and shrinking) battleground to keep x86 relevant, amid a more mobile, visual and power sensitive world.

What do you guys think? Can Intel do it?

Update: Peter N. Glaskowsky, a technology analyst for The Envisioneering Group calls this post preposterous over at C|Net. He argues that Intel can thrive, and that my idea of a post x86 world is wrong. Perhaps “post x86″ isn’t the most elegant way of summing up the trends of mobility and graphics, but I think it works. I’m also not counting Intel out when it comes to driving its x86 chips into newer markets (and say so in the comments), but I don’t think it’s “preposterous” to have a debate on this topic. Given the thorough response by Glaskowsky, those commenting on his post and the comments below, it’s something we should be talking about.

You’re subscribed! If you like, you can update your settings

  1. What does Intel have that nobody else does? A failed investment in WiMax, a lot of overhead, and a late start. Why would they succeed?

  2. I’ll tell you what Intel has that nobody else does.

    + Inhouse process development
    + Big R&D budget
    + Deep pockets
    + ITS OWN FABS

  3. Yetanothertechie Thursday, March 12, 2009

    It’s not hard to imagine Intel will eventually feature down Larrabee so that it can deliver an end-to-end graphics roadmap to complement its CPUs from server->embedded applications.

    Once that is done, what will Nvidia graphics pair with? Second-tier CPUs like Via.

    I think Intel is on track with whopping 55% gross margins to fund their investments of choice, and that the breadth and maturity of the x86-based software ecosystem will be enough to compel OEMs to use embedded Atom instead of other existing ARM-based SoC alternatives.

    This all depends on whether (1) the courts will observe Intel’s rights to refuse third-party graphics pairing with Intel’s CPUs and (2) third parties wait around for TSMC’s Atom SoC or move on with SnapDragon, OMAP, or Tegra.

  4. I’m skeptical. For one, Intel has never succeeded in a commodity market (they get out — see DRAM and communications). And that it what we’re looking at.

    So it’s not Intel vs ARM, it’s Intel versus Qualcomm, TI, Freescale, Samsung, etc — all chip makers used to living on much smaller margins. For mobile devices, battery life will always be important, and this is likely to continue to be a big edge for ARM for the foreseeable future.

    In the phone & internet tablet market, software compatibility won’t be important — since desktop software (and boot times) don’t work well on a smaller, tablet format. If it did, WinCE would’ve been a success (along with tablet PCs, too). But, as the iPhone shows, new approaches work much better.

    On the graphics side, I’m skeptical of Larrabee. Intel has never made a competitive GPU, and I’m skeptical that Larrabee will change that.

  5. Jesse Kopelman Thursday, March 12, 2009

    Larrabee is probably the key. If it is successful, then x86 has a shot at remaining relevant in the greater market. If it fails, so goes the x86-everywhere strategy.

  6. Stacey Higginbotham Thursday, March 12, 2009

    Tony, you’re right. It is Intel vs. the Arm licensees, especially Qualcomm, Nvidia’s Tegra and Texas Instruments. As for the commodity market point, that’s true, but Intel seems to be making a few different moves that are more customer-friendly like bringing out a video-playing version of Atom more quickly and its deal with TSMC.

  7. Patrick Moorhead Thursday, March 12, 2009

    I will just provide links, with no commentary.

    Business Week: “Is Your PC a Graphics Wimp?” http://www.businessweek.com/technology/content/aug2007/tc2007081_108723.htm?chan=search

    InformationWeek: “Intel Cites Graphics Problems In Centrino 2 Delay”
    http://www.informationweek.com/news/hardware/processors/showArticle.jhtml?articleID=208400736

    Notebooks.com: “AMD vs. Intel Integrated Graphics Video”
    http://www.notebooks.com/2008/08/20/amd-vs-intel-integrated-graphics-demo-video/
    The Inquirer: “Intel’s G965 embedded graphics stink – official”

  8. majortom1981 Friday, March 13, 2009

    Yes they wil lthrive. You need something to feed the gpu’s.Also intels xscale cpus are still fast even if they are not made any more .

    The touch pro has a VERY slow qualcom processor in it. Even though its rated at 52mhz5 (around there) . IF intel can do better marketing for its products they will do pretty good.

    Just look at the netbook craze . Intels atom processor is at the heart of most of them.

  9. I work for Intel (Stacey, you know how to reach many of us). Interesting post and debate. But, is anyone following the growth rates of laptops? Netbooks? Embedded computers? Are phones become more like small computers or visa versa? Look at the prices of PCs, laptops and chips the past 25 years. Don’t those price drops and Moore’s Law address, um, “commodity concerns?” As for WiMAX: LTE still needs to get things working, and ask investors to spend a ton of $$$ to get into the market while $$$ is being spent on 3G. But worse (!) case, even if WiMAX didn’t make it in a part of the world, LTE has moved up their schedule ~4 years already…that alone is a win-win for everyone. But – look at spectrum, not just press clips.

  10. Frank Miller Friday, March 13, 2009

    I’ve argued (against you ;) ) that the existing software base is the real driver for x86. The question I have is whether the applications that have been developed over time for x86 translate to the platforms that ARM is strong in. I obviously don’t know the answer, but I suspect the answer is yes and no. Some will and some won’t. Some apps just need more screen real estate, some just aren’t useful when you’re on the go. But there are lots that will. It can certainly be said that the argument for software momentum preserving x86 is less compelling in the mobile realm (and I’m referring to platforms smaller than netbooks here).

    I guess my opinion here is that it depends a lot on Intel. The Atom is a very nice device and certainly much less power hungry than anything else they’ve done, but its not a real competitor to ARM on these platforms. The netbook is about a small a platform as you can get to with Atoms practically, and it just so happens that there are lots of x86 that run well on that form factor. That’s why I think its taken over so quickly from the initial Linux implementations.

    If Intel wants x86 in phones or MIDs, it will have to do an implementation that is on par with what ARM does. What does that mean? First, it means they need an x86 that uses probably another order of magnitude less power per instruction than the Atom does. They also need to do real integration. The ARM devices that are winning now are those that have builtin memory and storage controllers, Ethernet, USB, IR, and yes, graphics. If there was a chip on the market that was x86 like the TI OMAP, ARM and Texas Instruments would need to be very afraid…

Comments have been disabled for this post