Windows Plus ARM Equals Trouble for Intel


Within the next two weeks, Microsoft (s msft) is expected to announce a version of Windows that will run on ARM-based (s armh) processors, the energy-efficient chips that currently power many of today’s smartphones. If reports from Bloomberg are accurate, the 25-year paradigm of Windows computers running mainly on Intel (s intc) processors is likely to be turned on its head, due largely to the need for devices that are more mobile and can run all day on a single battery charge.

Indeed, when Microsoft announced a licensing partnership with ARM in July of this year, we suggested one likely reason was to advance Microsoft’s efforts to port Windows to mobile processors. Why would the company want to do so when its platform runs perfectly fine on x86 chips, such as those from Intel and AMD (s amd)? Wireless technologies such as Wi-Fi and mobile broadband are enabling consumers and employees alike to connect away from the desktop, causing sales of pocketable smartphones and tablets to rise dramatically. Such devices are expected to begin outselling traditional computers within two years, and short of its new Window Phone 7 phones, Microsoft has no presence on any of these mobile machines.

If news is coming from Microsoft about Windows running essentially on smartphone chips, it’s highly likely to happen on Jan. 5, 2011 at the Consumer Electronics Show. Microsoft CEO Steve Ballmer is scheduled to take the stage for an evening keynote, providing a highly focused audience for an announcement. I’ll be among the crowd, attending live, and until then, I’ll be wondering what specifics Ballmer will mention, because this development doesn’t necessarily mean a full-blown desktop version of Windows will be made to run on processors that use the ARM architecture.

The current Windows desktop platform, Windows 7, is far too “bulky” to both exist and run on devices with limited storage capacity, memory and processing power. So it’s possible that a Windows “Lite” version could appear: a platform that provides the most used consumer features only, for example. Or Microsoft’s new Windows Phone 7 platform, which the company says has created 1.5 million device sales to carriers, could be extended with new features and be slated for tablet devices. I find that unlikely, and think the possibility exists that Ballmer will simply tout Windows Embedded Compact: the limited subset of Windows that was shown on tablets earlier this year at the Computex show.

While the specifics of any Ballmer announcement on Windows for ARM processors are up in the air, one thing is certain: Intel’s cash cow is slowly moving on to greener pastures. Much like Microsoft, Intel is being left out in the cold when it comes to mobile devices. Its chips are the powerhouse in the desktop and server processor markets, with an estimated 86 percent market share, but the future is in mobiles. That future looks still looks relatively bleak for Intel, as its Atom chips still consume too much energy compared to ARM-based processors, and its main operating system for mobile devices is Nokia’s (s nok) MeeGo platform: an unproven entity that will face Apple’s iOS (s aapl), Google’s Android (s goog) and other incumbents when it arrives in devices next year.

The Bloomberg report does keep the door cracked open for Windows on mobile devices for Intel, however, saying this:

“The new software also will be able to work on Intel Corp. and Advanced Micro Devices Inc. processors — the chips that power PC versions of Windows. While other versions of Microsoft software aimed at phones and mobile devices work on ARM chips, this is the first time it will make a full version of Windows available on that technology.”

If true, it represents the first time I can recall that a popular operating system will be able to run on both ARM and x86 architectures for mainstream consumers. That’s worth noting, because it means people can make an apples-to-apples comparison on Windows devices that run on Intel chips vs chips from many others, such as Samsung, Qualcomm (s qcom), Texas Instruments (s txn), and Nvidia (s nvda) to name a few.

Put another way: Intel will actually have to prove that its Atom efforts have made progress in the eyes of consumers, even as the next-generation ARM chips arrive next year with even more computing power and judicious battery life. The company better hurry, because ABI Research’s prediction from nearly a year ago is looking true: ARM-based mobile devices will surpass x86-based devices by 2013.

To be sure, the entire situation reminds me of a modern-day version of the game, “Musical Chairs.” For years, both Microsoft and Intel have paraded around, earning billions while the music played on. But neither have reacted well to the shift toward mobile, and the band is about to take a break. Depending on what Ballmer announces at the CES, Intel might not have a comfortable place to sit when this round of music ends.

Related content from GigaOM Pro (sub req’d):



mobile platforms are gonna question of Life & Death for every player in the software market. i see, most trouble of ms is not to port their software towards new chips, but simple hatred from their own users: many of’em have been too tired out of ms’s high prices & their lowest responsibility of security:-)


News of Microsoft’s ARM efforts are irrelevant to Intel.

The reason is that Microsoft has already lost the phone market. This has resulted in Microsoft also losing the tablet market.

So Microsoft is already in an unrecoverable position in tablets. The fact that Ballmer may do yet another keynote about another embarrassing effort in tablets will have no effect on Intel’s future.

The disastrous decisions that cost Microsoft the mobile/tablet market were made some years ago. That’s what really impacted Intel’s future, not next month’s hapless Ballmer keynote.


Is the market asking for Windows-on-ARM? or Microsoft feeling compelled to introduce this combination.

Windows-on-ARM is a tough sell as compared to the other OS-CPU combinations that exist.

Kevin C. Tofel

Looking at sales growth, the market is increasingly looking for capable mobile devices such as smartphones and tablets with a broad app ecosystem. That doesn’t necessarily mean Windows but Microsoft can’t afford a mobile future without a foot in the door.


“If true, it represents the first time I can recall that a popular operating system will be able to run on both ARM and x86 architectures for mainstream consumers”

um, ios?

Kevin C. Tofel

Great point – although the UI is different and the platforms are targeted for different devices across x86 and ARM, you’re right that iOS is derived from OS X. We can’t really make that apples-to-apples comparison I mentioned between chip architectures as a result, but excellent point!


As more mobile devices are using ARM, MS should jump on this as well. They already started with WP7 and need to keep the ball rolling to be able to be in the game. As far as Intel, they also need to come back with a new ARM processor to be able to continue as the big chip maker. If I remember correctly, Intel sold the ARM division? With Dual Core ARM coming out, with great battery life and great power to push the new Tablets to the extreme to get the mobile worker on the road connected and getting the job done. Things will change.

MS just needs to focus on keeping a touch UI OS for tablets and another one for the x86 Chip. Just look at Apple with the Mac and the iPad. Just need to do what the consumer and business users needs on a mobile device. Then they can work on adding new features you can do on a mobile device as they go on.

Kevin C. Tofel

Yup, Intel sold its XScale division to Marvell back in 2006, IIRC. Surely it had visions of the Atom line at that time (or something like it) but we’ll have to see if that pays off in the long run.

Shah Ullah

How do folks feel about the idea of Intel opensourcing Atom?

Let TI, Samsung, Qualcomm (with some separate issues around their strength in baseband withstanding) and other semi houses contribute to competition against ARM. It does them well to not have to pay royalties to ARM and ARM would not exist without the guys who actually do manufacture silicon. Also allows Intel to stay in the low-power game and maybe move into other areas of the value chain while they save new penetration.

Risk: As low-cost chips eventually become the norm for enduser devices and even starts to seriously chew into cloud infrastructure, Intel would have shot themselves in the foot. But again, if the strategy works and they have penetration, they can explore other areas of monetization.


Let’s hope they don’t relapse into a Windows Mobile-like product for mobile use.

James Kendrick

The problem with Windows apps on such a platform is not the recompiling, it is the user experience. Windows tablets have terrible tablet user experience now, largely due to apps not designed for such usage. This will continue to plague the Windows ecosystem, unless they use WP7 apps designed for such usage.

P. Douglas

I think MS should just extend Windows Phone 7 OS to tablets for the time being, and optionally phase it out when Windows 8 arrives. If MS is worried about revenue, I believe it should just charge $40+ dollars to license the product, and create a first class ecosystem to compete with the less expensive Android OS. Indications are that Windows 8 (or at least a version of it) will be built from the ground up for touch. The above approach gives MS a lot of time to get touch right in Windows 8, while having a product in the tablet market for the next two years.


You are right, but you are missing something. If there is a major shift from keyboard-oriented platforms to touch-oriented platforms, those applications, or to use the antiqued term, “programs,” have to make the transition as well or be left behind. It is in the best interest of software companies to understand and adapt to a touch interface device. Many of those companies will be more comfortable with Microsoft’s platforms, especially in corporate.


“25-year paradigm of Windows computers running mainly on Intel processors???”

Get your facts straight. Windows NT 3.1 ran on DEC Alpha, and MIPS as well. PowerPC was added to the supported hardware from Windows NT 3.51. All hardware platforms were supported up until Windows NT 4.0.

Porting Windows to ARM is simply an evolutional phase as ARM becomes widely used for mobile devices which were non-existent during the time of Windows NT.

Kevin C. Tofel

That may well be true (and I thank you for pointing it out), but by and large, WinTel is the elephant in the room in most cases today, which is why I said “running mainly” on Intel.


I am going to agree with the comment. Having watched ARM for over a decade, along with RISC and Acorn, it is hard not to equate MIPS and such with the evolution of this processor architecture. The multi-core ARM processor of the near future is neither just a mobile phenomenon, nor under-powered. It is a different architecture.

And, iOS is OS X. To miss this is to miss the future of the iPad and iPhone, and really will not put 10.7 and the launcher in perspective. Microsoft will follow a similarly reductive process in Windows. It will also require the redesign of Office.

Lucian Armasu

MS won’t put WP7 on tablets until they give Windows 8 light a run for it. If by 2014 WP7 is still around, that’s when they’ll probably use it when Win 8 light will fail – and it will fail. What’s the point of using a dumbed down Windows version that has NO backwards compatibility with previous apps. What is the benefit in that? In a way, Windows 8 for ARM will be a completely new OS, even if it borrows code from Windows 7.

Android and iOS will have huge ecosystems on tablets by the time Windows 8 for tablets gets released in 2013.


Microsoft can provide tools to recompile .exe apps to have them run on ARM Processors, just as most Ubuntu apps have been recompiled for ARM, they can also make some kind of Applications Marketplace for Windows, also integrating a bunch of security control stuff so people get to have less of these malwares.

I am sure Microsoft does have Windows 7 for ARM ready now. The question is do they want to open the bottle now or wait for later. Them opening up for ARM now would be considered an insult by Intel. But Intel started it first by working on Linux with Moblin and Meego.

Remember what happened when Adobe started working on ARM, with Flash support on all devices, that triggered Microsoft Silverlight.

When ex-partners do things against Microsoft, they are quick to launch their own counter attacks.

Lucian Armasu

Easier said than done. Isn’t making apps work on 64 bits Windows simpler than porting them to ARM? It will take many years till even some of the most popular Windows apps will be ported. Plus they’ll all need a revamped UI for the touch interface, too.

It was rumored that MS will encourage developers to make new apps as web apps for Win 8, probably so they don’t create a big fragmentation between Windows 7/XP and Windows 8 for ARM.

But then you have to ask yourself? If Windows 8 apps will be web apps, what’s going to stop people to use any other tablet OS?

Windows lock-in will be minimal, and people will have true choice between OS’s then. It will be especially hard for MS to enter the tablet market 2 years from now when Android and iOS will be already established OS’s in the tablet market, not much different from the WP7 story today.

Comments are closed.