Can the Wintel Marriage be Saved?
by Mark W. Hibben
A Mid-life Crisis
This past year has been a stressful one for the Microsoft/Intel business marriage, long dominant in personal computing. Although the marriage was always an “open” one, with either partner free to “see other platforms”, such liaisons hadn’t amounted to much. Intel had sold a few processors for Linux and Mac OS X systems, while Microsoft had stayed close to home and attended to its Windows knitting. Up until last year, both partners had been secure in the knowledge of the dominance of their Wintel collaboration.
Now the marriage faces a “mid-life crisis” as Windows PC sales stagnated in 2012, with consumer PC sales actually declining as consumer interest has turned to more exciting tablets and smart phones. As so often happens, insecurity has driven the partners apart, rather than brought them together. In 2011, Microsoft started chasing that cute little ARM processor (for Windows Phone and RT tablets), while Intel, not to be outdone, tried to attract Android interest in its own smartphone offering. Neither of these “affairs” have worked out particularly well.
As I discussed in Microsoft’s Mobile Crisis, the Windows on ARM initiatives (Windows RT, Surface RT and Windows Phone 8) aren’t going well. I admit hard numbers will be hard to come by until the earnings season, but I remain convinced that RT tablets and convertibles are being cannibalized by Windows 8 touchscreen devices, and that Windows Phone 8 will suffer the fate of its predecessor.
The reasons for the failure of the Windows on ARM strategy will be debated for a long time to come. Why has Microsoft been unable to do to its mobile competitors (specifically Google and Apple) what it did to Apple when Windows was first introduced in the mid 80’s for the desktop? Here are my reasons:
1) Weaker competition: In the mid 80’s, Microsoft faced a much weaker competitor in Apple than it now faces in either Apple or Google today. Apple is now the larger company with deeper pockets. Last quarter, Google had revenue of US $14 billion, Microsoft had revenue of US$ 16 B and Apple had revenue of $36 billion.
2) A loyal constituency: In the mid 80’s, Microsoft had a large constituency of DOS users who wanted to upgrade to Windows. Microsoft made sure of this by targeting the same Intel x86 processors for Windows that were running DOS. Microsoft didn’t try to create a version of Windows that would run on the Motorola 68K processors then used in Macintosh. No such base of support exists for Windows RT, and even Windows Phone 7 users can’t upgrade to WP 8. There was hope that Nokia would deliver its Symbian user base into the waiting arms of Microsoft, but that hasn’t happened, because there’s effectively no upgrade path. Symbian users just have to buy a new Windows Phone.
3) It’s the ecosystem: The ecosystems surrounding the platforms have become much more important than they were in the mid 80’s. Back then there was no Internet, no Cloud services, no social networking, no online shopping for apps, music or video. The Windows PC ecosystem mostly consisted of Microsoft, Intel, and various software and hardware developers. Today, both Apple and Android have much more developed ecosystems and a larger base of developers for their platforms as well. Both Android and iOS have much larger user bases, each with over 400 million users worldwide. Almost everyone knows someone with an Android or iOS device, so joining either group is very much a social act.
4) The rise of the custom processor: In my recent post, Texas Instruments and the Big Chip Maker Anachronism, I described how the role of the microprocessor maker had fundamentally changed for mobile devices. TI had been trying to be the Intel of mobile, but mobile device makers like Apple or Samsung don’t need an Intel. Apple started designing its own System on Chip (SOC) mobile processor for the first iPad, and having a custom-designed chip affords some advantage in optimizing battery life and system performance. Neither Microsoft nor most of its partners have the expertise to design ARM based SOCs, so they buy off-the-shelf processors from Qualcomm or nVidia.
Meanwhile, Intel’s manufacturing partners for Intel-based Android phones, Motorola, Lenovo, and India’s Lava are all refraining from bringing their phones to the US market. Enough said.
Renewing the Relationship
Now that Intel and Microsoft have had their “flings”, it’s time for these two to get back together. Reunited, they can compete far more effectively in Mobile than they could apart. So I’m not talking about abandoning the Mobile market, but I am talking about abandoning ARM. Wait, you say, doesn’t Mobile mean being on ARM?
Up until recently, that would have been true. It certainly was true when Microsoft first announced the Windows on ARM initiative at CES in January 2011. But Intel keeps pushing the technology envelope, and it may be able to triumph ultimately over ARM in the same way it did over PowerPC.
This year saw the advent of a new line of Intel processors (Ivy Bridge) based on Intel’s latest 22 nm process. The “22 nm” (nanometer) designates the size of typical features for transistors in the integrated circuit. In integrated circuits, the smaller the feature size, the better, especially in terms of power consumption. Microsoft’s Surface for Windows 8 Pro has an Ivy Bridge processor. Yeah, it’s a little heavier, and its battery won’t last as long as the ARM-based Surface RT, but it’s also a much more capable computer able to run full-fledged Windows 7 software.
Intel is also starting to roll out new ultra-low power Clover Trail Atom processors for use in tablets and smart phones. These appear to be on a par with ARM processors but are still being made on the older 32 nm process. The really good news for Wintel comes next year, when the next generation of Core processors comes out, code-named Haswell. Haswell will feature a new processor core architecture and be scalable from mobile device SOC’s to server processors. Haswell will also be built using the 22 nm process and thus will bring the 22 nm process to Intel’s mobile SOCs for smartphones and tablets.
At this point, Intel will be able to give ARM a run for its money, and working with Intel, Microsoft will be able to customize Haswell processors to meet its hardware needs. Imagine a true Pocket PC no larger than a smart phone that can run Windows 8 Pro and all your Windows 7 software, with video output to a high definition monitor. And it will be able to make calls, too.
Getting Back on Track
At the end of my last post, Microsoft’s Mobile Crisis, I left unsaid what I thought Microsoft should do about the crisis. Now, let me be specific. Microsoft needs to abandon its Windows on ARM initiatives, Windows RT and Windows Phone 8, in order to focus on building a better Wintel mobile experience. I’m convinced that Windows 8 on Intel touch screen devices will be a winner, but it still needs a lot of work. Integration of the Desktop and full screen modes especially could be better. As I pointed out in Windows 8: a Competitive Assessment, many Windows 7 users will be dismayed by their so-called upgrade to Windows 8, if they don’t have a touch-screen. Wintel can ill afford to antagonize their base of support.
Giving up on ARM is a retreat, I realize, but a vitally necessary defensive measure to protect the Wintel base of support of roughly 1.5 billion users. Once the base is secure, it will be possible carry on the fight for the “next billions” of mobile users from a position of strength. Do I expect Microsoft’s management to realize that Windows on ARM is a lost cause immediately? No, I don’t. Such realizations usually take a painfully long time, and in Microsoft’s case, will probably require regime change.