Apple On Intel, The Final Chapter (At Least For Now)
Brian Dipert - June 4, 2005
Continued from 'Apple On Intel, Part II'….
Back to the more likely scenario….Apple on Intel's existing CPU architectures. Painful? Yes, but not fatal. I happen to know, courtesy of my Intel connections, that both Intel and Apple have been running O/S builds on Intel-based hardware designs for years, as a just-in-case hedge. Folks often point out that OS X derives from the NEXTSTEP O/S that Jobs' team at NeXT developed, which itself derives from the Mach kernel and BSD implementation of Unix…..all of which optionally run (or at least ran, at some point in time) on x86 hardware. Look, though. OS X was announced in 2001, but Apple acquired NeXT in late 1996. OS X has, therefore, been in development for almost nine years. That's a whole lot of time and code beyond its x86 heritage. Still, it's possible. Oh, and by the way….I also happen to know that Steve Jobs keynoted an Intel Sales Conference a few years ago.
Next problem; device drivers. Apple-authored drivers are a recompile-and-tune away from migrating to a different CPU (how simple it all sounds in theory!). Apple, like Microsoft, also bundles a huge suite of third-party drivers with its operating systems. I don't know what percentage of them are written by engineers in Cupertino, versus by the third parties themselves. But, assuming Apple has access to the source code in either case, it's simply a matter of enough engineers, enough time, and (to motivate third parties who control their own source code) enough money; it's doable. Particularly because the number of peripherals Apple needs to support is a fraction of what's in the Windows ecosystem.
Final problem; applications. For Apple-authored apps (iLife, iWork, Final Cut Studio, Logic Pro, the 'Express' products, etc) see above re. Apple-authored drivers. Definitely doable. But what about third-party apps? This is why, logically, Jobs makes the announcement at WWDC, and delays the hardware rollout for at least a year (watch out for the Osborne Effect, though!). For a while, at least, companies will need to sell two or three (see earlier Itanium prediction) different compilations of their code. This won't last long; Apple will be motivated to 'encourage' customers to upgrade their hardware, and multiple line items are a pain in the bum for customers who don't know or care what CPU's in their box. Microsoft learned this with Windows CE; it's why they dropped MIPS and SH CPUs and now only support ARM-based processors. And for the slow-to-migrate (or no-migrate) apps, hardware-emulating software layers can always provide a bridge, albeit a somewhat performance-sapping one.
Next question; does this mean Apple's going full-bore into the software business, selling OS X and its apps against Windows (and Microsoft's apps for Windows) to run on generic x86-based hardware? I very much think not. Fundamentally Apple, unlike Microsoft, is first and foremost a hardware company. And support issues exponentially increase as the number of hardware and software combinations you need to comprehend increases. This is why OS X is a more stable experience than Windows; the solution set Apple needs to grapple with is much, much smaller.
So who wins if the Apple-on-Intel predictions pan out? Intel picks up another few percentage points of market share. On the one hand….no big deal. On the other, it benefits from the prestige of being anointed by Apple, whose brand awareness and esteem are paramount. And Apple? High performance. Low power consumption. Low cost and (related) abundant supply, courtesy of the engine that drives over 80% of the world's PCs (but, I might add, none of the world's next-generation game consoles…sorry, Infinium Labs, you don't count). Seems like a win to me, if they've got the guts to roll the dice and make the vision happen, even considering the short-term transition pain.
Losers? First off, AMD. Two words why they didn't get the Apple nod; manufacturing capacity. Plus, notebooks are the fastest-growing segment of the computer biz, and AMD doesn't have an answer to Intel's Pentium M in the thin-and-light space where Apple's iBooks and PowerBooks play. Freescale and IBM lose too, at least from a perception standpoint. IBM more so; Freescale's given strong signals for a while now that they're more interested in focusing their PowerPC efforts in the embedded space. In spite of snagging all three next-generation game consoles (Microsoft's Xbox 360, Nintendo's Revolution and Sony's PlayStation 3), IBM's been burned by the same 'Apple prestige' factor that'll boost Intel's fortunes.
I'm scheduled to attend Jobs' keynote Monday morning. Check back afterwards for more ramblings. Until then, I welcome your thoughts!