10 Comments

Summary:

Intel’s Atom chips are finally ready for smartphones and we should see the fruits of Intel’s labor soon, just not this year. Instead, trade shows in early 2011 will set the stage for Intel smartphones. The problem is, consumers really don’t care who powers their handset.

Intel has talked about its plans to push its chips into smartphones for well over a year. It appears such goals won’t be realized anytime soon. Handsets powered by Intel chips are expected to debut in numbers at either the Consumer Electronics Show or the Mobile World Congress events, both of which take place in early 2011. But such devices would likely be prototypes or used to announce design wins rather than actual shipping products. Intel CTO Justin Rattner tells Wired that these two trade shows are the “window of opportunity” for Intel to power devices in the fast-growing smartphone segment. I agree with Rattner that it’s good timing for Intel, but I still think the company faces a huge challenge in this market.

In fairness, Intel has made great strides with its Atom platform. Originally, such chips were used to power Ultra Mobile PCs and netbooks, but Intel’s Moorestown chipset — part of the Atom line — is meant to run inside smartphones. That’s traditionally a market ruled by processors built on ARM architecture, not Intel’s x86 platform. While ARM chips aren’t more powerful than those of Intel on a relative basis, they have offered better performance per watt of power used — in other words, ARM devices tend to run far longer on a single battery charge. But the latest Intel Atom chips are now approaching ARM-levels of power efficiency.

Unfortunately, Intel has a number of hurdles to overcome — simply building an chip that sips battery power instead of guzzling it is no guarantee of success. For this reason and others, Om has deemed Intel a mobile loser. I agree and think there are additional challenges that will limit Intel’s success in smartphones, the most important one being: why do consumers even need an x86 chip inside their smartphone? Support for the x86 platforms such as Windows, OS X and desktop versions of Linux aren’t required by the smartphones of today — or tomorrow, for that matter — when there are very capable mobile platforms that can run on ARM devices.

The fact is that most consumers don’t know or care whose chip is inside their handset. I’m willing to bet that if I asked ten random people what the chip is inside their smartphone, perhaps one or possibly two could provide an accurate answer. And of those that could, I’d further bet they’re as geeky as me and like to know about the guts of their devices. Most other consumers only care about what they can do with the device, not what brand of chip is inside. Yet Intel is clinging to its old “Intel Inside” computing strategy in a smartphone market. That marketing approach simply won’t work in today’s world.

To this point, there’s no compelling evidence that consumers even want to run a full-blown x86 desktop operating system on a mobile device. Instead, people are embracing lightweight operating systems supplemented by powerful mobile applications. This is a lesson Apple has learned and successfully implemented with the iPad — the device might fully replace a computer, but for over 3 million buyers in 80 days, the ARM-powered device is useful and provides a pleasant, but still productive, mobile experience.

Ironically, software may be the biggest challenge of all for Intel’s smartphone hardware plans:  you need a solid operating system for your chips to run. All of the currently popular mobile platforms are have adopted and been optimized for ARM chips — Apple’s iOS4, Google’s Android, Research in Motion’s BlackBerry and Nokia’s Symbian all run on ARM, not x86. Intel is reportedly working on a port of Android for x86 and I’ve even installed Android myself on a portable  x86 computer, but it will take time for Intel to get Android optimized.

Intel also has the MeeGo operating system for Atom in the works — which is derived from its own Moblin Project and Nokia’s Maemo system — but will handset makers want to use it? Not even partner Nokia has announced any Intel design wins for a smartphone. And since Meego is still in the very early stages of development — the first bare-bones build with an interface arrived just two days ago — it’s difficult to say when the platform will be competitive with iOS4 or Android.

With so many viable platforms available — and more importantly, tens of thousands of applications for them — what’s the compelling reason to use Intel over ARM? I suppose we’ll see if there is one at CES or MWC in early 2011, but until then, it’s an ARM world.

Related content on GigaOM Pro (subscription required):

For Phones, The Future Is Multiple Cores

  1. Jacob Varghese Friday, July 2, 2010

    I don’t think companies want to be tied to just one supplier.
    1. Look at all the supply issues around Intel’s latest laptop processors i3, i5, and i7.
    2. Restrictions on motherboard design that prevent Nvidia from building their graphics chips into these boards.

    ARM is the way to go. Companies can work with a variety of suppliers for chips – Qualcomm, Nvidia, Samsung….

    Share
  2. Enjoyable analysis.

    You say: “Support for the x86 platforms such as Windows, OS X and desktop versions of Linux aren’t required by the smartphones of today — or tomorrow, for that matter — when there are very capable mobile platforms that can run on ARM devices.”

    The odd man out is Windows. OS X and Linux already can run on ARM. (This ignores the question of whether a desktop experience is appropriate for a mobile device, but I can imagine an iPad+ where a similiar UX might work.)

    Share
  3. Feels like de ja vue, we are back to the old PC compatible days inly now for smart phones.

    A little competition on the OS is good, a lot will fragment development – like the whole flash thing on Apple.

    Share
  4. X86 is the most preferred platform for developers around the world and the number of applications that exists for X86 platform is at least a thousand times more than any ARM platform. The user experience for an X86 based smartphone will be far greater than that of an ARM based smartphone as the user can almost download any application of his choice from the internet and run it on the phone. Its going to be almost a PC in you pocket.
    The best part is going to be – application development, unlike ARM you dont need to cross compile and deal with compatibility issues. Instead take any source code(from millions of open source projects) , switch library (mostly GUI), compile and run!

    Share
  5. as strictly a hackers toy it will be a bit of fun to install windows on one of these.

    Share
  6. The usual affluent but numerically tiny Apple crowd can afford to own multiple Apple devices, so to these people, they have the lightweight iPhone operating system on their iPhones and iPads, and powerful OSX Mac desktops.

    For the masses of average folks though, it’s their featurephone, and their Windows netbook- 30 million sold in 2009, or laptop or desktop. With a full-blown x86 desktop operating system on a mobile device, the average folks may just go only with a mobile device.

    Share
  7. [...] The huge need for mobile processors isn’t lost on companies like Intel, which until only a few years ago focused its x86 chip architecture on fast CPUs for desktops and laptops. With the mobile revolution underway, Intel is attempting to gain a foothold with lower-power Atom chips that sip power from the smaller batteries of mobile devices. But chips based on ARM’s architecture have always been power efficient and with each new iteration gain faster performance. As a result, handset makers continue to use ARM-based CPUs from Qualcomm, Samsung and Texas Intstruments, or, as with Apple and its A4 CPU for iOS devices, are creating their own processors. Intel’s late to the table, and Om thinks the company will be a mobile loser. I tend to agree. [...]

    Share
  8. [...] in the processor market, with Texas Instruments and Samsung in the mix, Nvidia still pushing and Intel pressing to get into the market. The ARM-licensing model means that an x86 style domination would be more difficult to maintain. [...]

    Share
  9. [...] This year may be remembered by history as the debut of the highly-capable ”superphone” but if Qualcomm has any say, 2011 could bring another quantum leap in smartphone computing, reports AnandTech. The mobile chip designer introduced plans for next generation silicon at an analyst event yesterday, claiming next year’s chips will offer five times greater performance with 75 percent less power draw  compared to its current Snapdragon line. Attributes like this will only further that gap between ARM-based chips and those of Intel, which has yet to find a foothold in the fast-growing smartphone market. [...]

    Share
  10. The point I think the article mises is that developing software for x86 is a non-issue. EVERYTHING runs on x86 already.

    The main hurdle in porting current x86 OS’s to a phone is going to be power consumption, but there are a number of very low overhead x86 OS’s available currently. Added advanced power saving features from organizations like Ubuntu and Microsoft are likely to appear if atoms ever do hit the handset market.

    I think the main hurdle here is going to be for ARM OS’s. Will there be enough development in meaningful applications for the ARM platforms to provide applications capable of competing with those currently available on a x86 desktop.

    Share

Comments have been disabled for this post