Wintel is dead or at least off the desktop. It became clear last week that Windows 8 on tablets will be essentially an ARM-only affair, putting the biggest potential tablet market squarely in the hands of ARM processor chips. The largest actual market for tablets -- Apple's iPad -- is already based on ARM chips, as is the first runner-up, those using Google's Android. Oh, the Research in Motion BlackBerry PlayBooks use ARM chips as well, not that many people have bought those. Ditto for Hewlett-Packard's short-lived TouchPads.
But Microsoft's move threatens to seal Intel's coffin. Although CEO Seteve Ballmer said Microsoft's Windows strategy was "Intel and ARM, not Intel or ARM," that's a misleading statement. Microsoft has since clarified the real strategy: ARM is for tablets that run Windows 8's new style of Metro apps, and Intel is for PCs that run both pieces of Windows 8: its modern Metro half and its legacy Windows 7 half. That means mobile is ARM territory across the board.
[ Read InfoWorld's coverage of Microsoft's Windows 8 grand reveal of its new UI and capabilities. | Take a visual tour of Windows 8's mobile-inspired Metro user interface. | Keep up on key mobile developments and insights via Twitter and with the Mobile Edge blog and Mobilize newsletter. ]
Yes, theoretically, tablet makers could use Intel's Atom chips, which powered the briefly exciting netbook market a few years back. But you see hardly any Atom-based tablets these days, just a few of the old-style Windows XP/Vista/7 tablets adopted in small part by government agencies but no one else. That's true even though Intel a year ago ported the smartphone version of Android to run on such tablets. (There is no Intel version of the tablet version of Android.)
Wait, you say: Google last week announced that Android 4 "Ice Cream Sandwich" (the unified tablet-and-smartphone version due by Christmas) and all future versions of Android would be optimized for both ARM and Intel chips. That should mean Intel-based Android tablets are on the horizon.
But let's remember why Microsoft said its strategy assumed ARM tablets and Intel PCs: Intel chips consume way too much power, resulting in unacceptably low battery life in tablets and smartphones. If Intel had a real answer to that issue, you know Microsoft would've been all over it. It could have planned for Intel-based tablets that ran the whole Windows 8, not just the Metro part. When you consider that Windows 8 is not likely to be released for a year or so, you can see Microsoft had no confidence that Intel's chips would be mobile-savvy in the near future.
Microsoft isn't alone in reaching this conclusion. Apple threatened Intel last spring to stop using Intel chips in its MacBook laptops -- the company's biggest seller -- because of their power wastage. Intel publicly acknowledged Apple's threat and said it would do better. Apple's threat carries real weight: Apple has transitioned its OS twice acrosss chip architectures, in a way that apps did not need to be recompiled to run. First, it moved from the Motorola 68000's complex-instruction-set computing (CISC) architecture to the IBM PowerPC's partially reduced-instruction-set computing (RISC) architectire, and then from PowerPC to Intel's x86 CISC architecture. Plus, Apple has already ported the core parts of Mac OS X shared with the iPad's and iPhone's iOS to the RISC architecture used in its ARM-based A4 and A5 chips.
Microsoft couldn't make the same threat on behalf of the PC industry, as it doesn't seem to know how to port Windows 7's core to RISC chips -- if it did, I believe it would have done so as part of Windows 8's ARM support. Though to be fair, in the late 1990s, Microsoft had ported the original Windows NT (the basis of today's Windows 7) to the PowerPC chip, though apps had to be recompiled to work on the PowerPC version, so perhaps it could do it again if it really wanted to.