MacPrices.net

Apple Switch To In-House Processors For The Mac? — Not A Good Idea

Wednesday, February 10, 2010

by Charles W. Moore

There’s been speculation since the iPad announcement that if Apple can make its own iPad silicon, why not in-house CPUs for the Mac as well? I have no idea whether Apple has been even remotely considering making such a move, but if they were to drop MacIntel in favor of homegrown chips, I don’t think it would go well for them, and they should tread very cautiously.

I have to admit that I was among the many who initially reacted with shock and skepticism when Steve Jobs announced at the 2005 Worldwide Developers Conference that Apple would be switching the Mac from PowerPC processors made by Freescale Semiconductor and IBM to Intel silicon.

After all, Intel had been the adversary — and sometime object of ridicule (if you’re old enough, perhaps you remember the “Toasted” ads and the bunny suits for example) during my entire tenure on the Mac up to that point. Intel was kind of like “The Others” to the “Lost” castaways.

And the PowerPCs RISC architecture was engineering-wise superior to the CISC technology used by Intel for its x86 series of microprocessors wasn’t it? (Actually, it turns out that there was more commonality than hard-core RISC partisans preferred to concede). Then there was the greater elegance of the whole PowerPC of idea/vision/concept/whatever — at least perceptually.

Ad: Compare MacBook prices at Mac Prices. Current models are available today for up to $250 off MSRP! Award-winning price trackers updated each morning.

However, I now freely concede that my reflexive reaction was dead wrong. Both of Apple’s “AIM” (Apple/IBM/Motorola — the latter from which Freescale hived off) had quite evidently lost enthusiasm for keeping pace with Intel technology in their PowerPC personal computer product development. G4 was pretty much stalled and moribund, and the G5, while reasonably powerful, was a hot-runner literally as well as figuratively — too hot for practical application in a laptop form factor, which by 2005 had become Apple’s best-selling computer system product category. In 2020 hindsight, it would have been disastrous — insane really — two persevere with PowerPC and not make the switch to Intel, which had emphatically not lost interest in personal computer chip development and innovation.

Subsequent history over the past four years pretty much speaks for itself. Following the Intel switch, the Macintosh platform went from languishing in the 2 to 3% North American domestic market share range to roughly four times that today, and to all intents and purposes virtually owns the $1000 and greater personal computer category

BetaNews’s Joe Wilcox reported last week that according to a source at one of the major market research firms, Apple’s US retail unit share has also now doubled from 5 percent to 10 percent for PCs selling between $500 and $1,000. Even more startling, says Wilcox, Apple increased its unit share from 79 percent to 90 percent in the market for “premium” PCs, meaning those selling for more than $1,000, or stated another way: nine out of 10 premium PCs purchased from US retail brick-and-mortar stores or online sites (including major chains and Apple Store) during the company’s fourth quarter was a Mac.

Would anything even close to this level of success and growth have happened had Apple stuck with PowerPC? I don’t believe it for a moment.

However, Forbes’ Lee Gomes wrote last week that that the future trajectory of the semiconductor industry is to some degree riding on the success of the iPad, whose A4 microprocessor “system-on-a-chip” silicon was partly designed in-house by Apple, most likely using expertise acquired via the company’s 2008 acquisition of Silicon Valley start-up PA Semi. Gomes thinks that if the iPad proves to be a marketing hit it might encourage Apple to consider making its own CPUs for Mac systems as well and thus could forgo paying the so-called “Intel tax.”

While it’s an intriguing line of speculation, Apple would want to proceed very cautiously, given that a major consideration would have to be continued compatibility with the Windows universe, which has been facilitated greatly by Apple’s adoption of Intel CPUs and motherboard architecture in its computer systems, which as we’ve noted not coincidentally coincided with the dramatic reversal in Apple’s market performance in terms of Mac systems sales, which has been largely driven by laptop sales.

Apple Matters’ Chris Seibold also thinks arguments in favor of going in-house with Mac CPUs would be abundant, including more control, more profits, and a chip expressly designed for Macs, all of which may seem good at face value but would be a strategy that could backfire royally if the Mac’s surge in popularity over the past four years really has, as I deduce, been based substantially on Intel X86 compatibility.

ZNet’s Adrian Kingsley-Hughes thinks it’s not so much “Intel Inside” that’s selling so many more Macs these days, but more to do with the Windows compatibility that comes with Intel silicon.

I’m inclined to agree more with both Kingsley-Hughes. Windows compatibility comes with Intel compatibility, and it it pretty difficult to rationalize away the correlation between the ability to run Windows natively on the Mac with excellent performance (or as good as you get with Windows) and an approximate quadrupling of Apple’s share of the PC market since 2006.

Better to focus in-house chip development efforts on the iPhone, iPod and soon-to-be iPad universe, where Apple is already dominant than to risk the Mac’s continued ascendency in the PC orbit by dropping Intel inside.

I think Apple could probably make its own chips to power the Mac, but the operative question here is would it really want to? I’m highly doubtful.

CM