Brian's Brain

In September of 2005, I (for the first time) called it. And an hour ago, Nokia actualized it. The x86 architecture has just established a hefty beachhead in its invasion of the ARM Empire. And with a portion of Nokia's future product line territory conquered, Intel partnerships with other mobile computing and communications companies will undoubtedly follow in short order.

Strictly speaking, this isn't Intel's first notable announcement in the mobile communications-and-computing product category. But with all due respect to LG, an embrace by the world's largest cellular handset supplier trumps the MID (mobile Internet device) news from 3GSM Mobile World Congress in mid-February. Neither Anand Chandrasekher (Intel senior vice president and general manager of the Ultra Mobility Group) nor Kai Öistämö (Nokia's Executive Vice President of Devices) was willing to discuss specific product plans, but of course I have ideas.

Öistämö was careful to state that the agreement was non-exclusive, and that it would have no material impact on his company's relationship with various ARM licensees. While in a sense I understand what he means, I think that in the long term he's underestimating the impact to ARM and its partners. To be clear, I don't see Atom descendents displacing ARM in traditional handset form factors and voice-dominant feature set packages. Keeping in mind Drucker's Law ("An emerging technology must be 10x better than the incumbent approach in order to have a reasonable likelihood of replacing that incumbent"), Atom won't satisfy the necessary criteria.

However, as anyone who owns an iPhone or other smartphone already knows, handheld and wirelessly connected devices are in the process of radically expanding beyond their voice-centric roots. Just yesterday I installed the Facebook app on my Windows Mobile handset. In recent weeks, I've talked about Twitter clients for various mobile operating systems. Between social networking sites and more general email, chat and web browsing functions, I don't think I'm alone in saying that voice calls comprise a slim percentage of my overall cell 'phone' usage profile nowadays. And I doubt this is a North America-only phenomenon; if anything, other areas of the globe with more robust wireless broadband infrastructures (i.e. Europe and Asia) are the trendsetters here.

So what form factors am I talking about? In this regard, Nokia's a natural partner. Take a look at my February, 2008 coverage of Nokia's Internet Tablet series, based on Linux and notably not including cellular voice and data subsystems:

Next, have a gander at my more recent writeup on Nokia's N95, which is Symbian O/S-based and cellular-augmented:

Now consider the just-started-shipping N97 follow-on to the N95, along with the rumored (and cellular-inclusive) N900 follow-on to the N810. This, I strongly suspect, is where Atom is notably headed, wherein its compatibility with abundant legacy code and development tools (therefore tapping into a legion of developers) give it an arm up over ARM (pun intended). And given the points made in the previous paragraph, this is where the bulk of handset sales will be in the future, as Danger foreshadowed a long time ago.

Given the Symbian-vs-Linux disparity mentioned above, it's no surprise that the second 'leg' of today's three-part partnership announcement involves Linux-based operating systems, specifically Intel's Moblin (co-developed with Novell) and Nokia's Maemo. I'll have more to discuss about mobile operating systems in a dedicated post soon to come. For now though, I'll devote the remainder of your attention to the third major tenet of today's announcement, that being Intel's license of Nokia's HSPA 3G cellular data technologies.

This isn't the first time the two companies have partnered on HSPA. Back in October of 2006, an announcement at the Intel Developer Forum unveiled a co-developed mini-PCI module for notebooks and netbooks...which never ended up going into production. Öistämö explained away this past 'speed bump' by noting that although the inevitable convergence of communications and computing had been conceptually obvious several years ago, the reality behind the concept took longer than originally forecasted to come to pass. But interestingly, there was no corresponding announcement today of a Nokia license of Intel's WiMAX work.

Chandrasekher once again expressed strong Intel support for WiMAX, but it's hard to not see today's developments as a back-step in the company's embrace of the technology (thereby making room for a pragmatic adoption of the 3G UMTS/HSPA...and presumably also future 4G LTE, although nothing was announced here today...cellular data alternative). Ironically, Intel and Nokia also had a past partnership in WiMAX that didn't bear fruit...Nokia planned to use Intel silicon in a WiMAX-augmented N810 Internet Tablet variant, and I've even seen prototypes with my very own eyes, but at least to date it hasn't gone into production.

WiMAX has seen relatively enthusiastic adoption in some global markets, particularly in emerging economies that view it as an economical and rapid-rollout alternative to wired (coax, copper, and/or fiber) broadband alternatives. But the U.S. GSM cellular providers are uniformly behind HSPA, with historical CDMA stalwart Verizon even joining them for the next-generation LTE rollout. Sprint/Clearwire owns the only notable WiMAX footprint here, and European cellular providers similarly have no notable WiMAX plans. At the end of the day for Intel, it's all about wireless mobility beyond today's limited-range and provider-fragmented Wi-Fi, and whether WiMAX or HSPA-then-LTE is the vehicle by which this objective is delivered is a secondary concern.