One of the most anticipated technology events in 2012 will be the commercial release of Microsoft's Windows 8. Microsoft not only wants to demonstrate their commitment to the Windows PC platform but also show that they can provide deep levels of innovation in new platforms like tablets and convertibles. Microsoft must also demonstrate to many that they can compete on many fronts with Apple, Google and Amazon on platforms, apps and services. Microsoft turned over every stone possible in their planning stage, eventually deciding to make a very huge hardware change in supporting the ARM processor architecture, a major departure from PC Windows platform exclusivity with X86, namely Intel and AMD. With risk comes return, but only if one mitigates the risk. What's becoming increasingly clear is that Windows 8 on ARM architecture, aka WoA, will face many hurdles which will need to be quickly overcome to successfully position it. If not, it will be very challenging for Microsoft, ARM, Nvidia, Qualcomm, Texas Instruments, their customers and their retail channels.
BUILD Conference and the Developer Preview
Let me start by giving a little background starting at Microsoft's BUILD conference. I personally attended Microsoft's developer conference, called BUILD, last September. It was one of the most informative and exciting Microsoft events I have ever attended, and I have been to many in the last 20 years as an OEM customer and as a hardware partner. Microsoft communicated with confidence and clarity a level of detail I had never seen before, which really gave me confidence that Windows 8 was headed in the right direction. I have loaded Windows 8 Developer Preview on almost 10 X86 PCs and so far I am very impressed with build quality, speed, and compatibility. I have heard the same from many respected name in the industry. There was no Developer Preview for ARM SOCs that attendees could take home with them, and there are good reasons for it.
ARM Incompatibility with Legacy X86 Applications
By definition, programs that have already been developed for Intel and AMD X86 PC platforms will not run on any ARM SOC (System on a Chip) from Nvidia, Qualcomm, Texas Instruments, or any other ARM licensee. The reverse is true too; apps written for ARM will not run natively on X86 processors. Software programmers are ingenious at emulation, but nothing has been announced about ARM emulating X86 nor do I think we will see it anytime soon. This is primarily because emulation eats up significant hardware resources which no current ARM-based SOC has to spare. Intel, on the other hand, is emulating ARM on their X86 Medfield platform as they showed at CES 2012. Obviously Intel believes they do have sufficient resources to do this.
So what does this mean? That Office 2010 someone bought for $120-350 or that Sims 3 Pets game you bought for $29 in September of 2012 will not run on a ARM-based Windows 8 tablet or notebook. This doesn't mean that WoA platforms won't have applications; they will. Microsoft has gone great pains to develop a new application development environment where developers can write new Metro Windows 8-based applications using a new and slick Visual Studio developer app. That doesn't change the fact that none of the legacy apps will work.
I've heard the argument that none of this matters because consumers only care about new apps and that they don't care about old apps. Examples like the iPad and Kindle Fire are given. The big difference is Windows, and what it infers. Windows stands for compatibility and being able to use legacy peripherals and apps. While I can probably count the number of $350 iOS apps on my toes, many exist for Windows. Some won't care, but for the buyers who are OK throwing out their old software library and starting over will buy a Windows 8 device with an Intel or AMD X86 processor.
Lack of Windows on ARM Desktop Means Lack of "Rich" Applications
Windows 8 for X86 has two operating modes, "Metro" and "Desktop" mode. Metro is the tiled based interface designed
The biggest implication of this is that WoA systems will not be able to develop what Microsoft calls "rich" apps, or those
I've heard the argument that says, "rich" apps are yesterday and the apps of the future are like Metro. To some extent they are right, but if you were around when the market moved from DOS to Windows, you know if took years to make the switch. Many enterprises still use DOS-based apps in a shell today. Just like many consumers fought moving from DOS to Windows, many will fight moving from Desktop to Metro. There are differences between then and now. The biggest difference is that on WoA systems, users won't be given both, they will only be given one choice, Metro. Secondly, there isn't a good way to make a dense layout work on Metro. Some will be OK with it, some won't. Those who won't will buy systems based on Intel or AMD.
Windows on ARM Peripheral Incompatibility
As I said before, backward compatibility has been a hallmark of Windows for decades, both on the app and peripheral side. I took for granted when I used my new Windows 7 system in 2009 that every one of my older mice, printers, joysticks, keyboards, USB video cameras, digital cameras, SD cards, USB memory sticks and scanners worked with it. That happened because Microsoft and its ecosystem invested millions of man-months into testing, validating, and in many cases rewriting or tweaking their drivers for the devices.
The stage has been set where the ecosystem will need to test and validate twice, once for Windows on X86 and Windows on ARM, and in some cases rewrite or tweak drivers. With peripheral margins as low as they are and without a recurring revenue stream from legacy peripherals, what is the big motivation to fix a broken driver? We will get a good sense of which peripherals work and those that don't as soon as the Windows 8 Beta is released next month. Then it will be up to the peripheral makers to decide if they will fix a broken driver of an old, and in many cases, discontinued peripheral.
One example really brings it to life for me. What will be the reaction when a consumer buys a WoA device, sets it up at home and one of their important peripherals doesn’t work like they expect? They will return the WoA device to the retailer and probably buy a Windows on X86 device.
The argument I hear on this objection is that Apple didn't have an issue with the iPad. That is correct, but the situation is very different. First, consumers expect any system device called "Windows" to be able to use their mouse, keyboard, printer, scanner, joystick, digital cameras and every other peripheral they used on their other Windows computer. That's one of the power of Windows that few expected from an iPad. The iPad was a new platform that didn't use a keyboard or a mouse, it used direct touch. It didn't have any USB ports or SD card slots and therefore most consumers knew it wasn't a PC and didn't have those same expectations. I suppose Microsoft could decide not to call this new operating system Windows, but I would bet 100% that they will. Therefore, consumers will expect compatibility.
Lack of Windows on ARM Design and Marketing Funds
Intel and AMD have invested billions (yes billions) of dollars into funding OEM and ODM designs and providing market development funds to OEMs, ODMs and distribution channels like Best Buy, Wal-Mart and even Amazon. Intel and AMD pay for all those TV, retail, web banner ads, and keyword searches that link a PC OEM to an Intel or AMD brand. ARM and ARM-based Nvidia, Qualcomm, and Texas Instrument do not and don't plan to as it's not built into their margin structures or their DNA. If they do some small token, it's significantly less in aggregate. Even if the ARM and the ARM ecosystem did decide to invest the dollars, they don't have the people to manage these programs at the ODMs, OEMs or retailers. Microsoft could decide to cover those expenses I suppose but I don't think Microsoft thinks it's their responsibility to do that.
Possible Depositioning of Windows 8 on ARM SKUs
When I connect the potential Windows 8 on ARM challenges above with the lack of marketing funds and resources, I see a possible scenario where Windows 8 on ARM SKUs are depositioned from the start. I also wouldn't be surprised if Intel and AMD took some of their marketing funds and resources and started to painting the following picture with OEMs, ODMs, retailers, enterprise buyers and even consumers:
- "Only Windows 8 on Intel/AMD lets you run Metro apps and Windows 7 legacy apps. No need to throw away all that money you spent on those apps. Windows 8 on ARM/Nvidia/Qualcomm/Texas Instruments will create a nightmare scenario where customer will return their tablets and notebooks because they realize that last month's software won't work with that new purchase. Don't risk burning that customer relationship or money on the product return."
- "Only Windows 8 on Intel/AMD will give you peace of mind on returns because peripherals will work well. Windows on X86 has been like this for 20 years. Windows 8 on ARM/Nvidia/Qualcomm/Texas Instruments will create a nightmare scenario where customer will return their tablets and notebooks because their peripherals won't work. Remember the Linux-based netbooks with 50%+ return rates? Don't risk burning that customer relationship or money on the product return”
- "Only Windows 8 on Intel/AMD lets you run new Metro apps and new "rich" desktop apps. You get the FULL experience, not the limited experience. Windows 8 on ARM/Nvidia/Qualcomm/Texas Instruments limits the apps you can run. Why settle for a limited experience?"
- "Think of Windows 8 on ARM/Nvidia/Qualcomm/Texas Instruments as the 'light' version and consider Windows on Intel/AMD the 'premium' version. Sure, you get an extra hour of battery life on the ARM/Nvidia/Qualcomm/Texas Instruments tablet, but is it worth risking all that functionality and compatability."
What I did above is a simple messaging thought exercise. Neither Intel nor AMD is going to allow anyone to enter their world defined as Windows PC platform without a serious fight. Intel and AMD will leverage core competencies, resources and 20 years of Windows experience against areas where neither ARM, Qualcomm nor Texas Instruments have. Nvidia is the exception in the ARM camp in that they have fought with Intel and AMD and have a lot of experience at it.
Windows 8 on ARM Challenge Ahead
With close to 100% market share, ARM and its ecosystem has had an incredible run in smartphones and tablets, two markets that are still growing much faster than Windows PCs. They have Apple to thank for building the iOS ecosystem and their three breakthrough platforms; iPod, iPhone and iPad. It's incredible to think now that ARM, Nvidia, Qualcomm and Texas Instruments will start to enter the Windows PC platform world now dominated almost 20 years by Intel and AMD. The potential threat of Windows 8 on ARM getting depositioned at the start is high, but could be overcome with time and investment. Intel and AMD will spend a lot of money and resources to educate the entire ecosystem and consumers to their perceived and real advantages. The question now is, what will ARM, Nvidia, Qualcomm and Texas Instruments do in response? Do they deprioritize the Windows 8 on ARM platform initially, focus on smartphones and Android tablets and when the situation is clearer, re-focus on Windows 8 later? Or do they address the challenges right now and start planning the counter-attack to Intel and AMD in their own Windows PC space?
