More thoughts on the Apple-Intel deal 


Well, it's a week since Apple's announcement that they will be switching to Intel processors from 2006 and the Mac community is starting to come to terms with what this will mean for the future of the platform. Overall, there seems to be an air of cautious optimism as the move looks likely to yield faster, cheaper Macs in the not-too-distant future. My own feelings have moderated from the initial shock-horror to mild puzzlement, although after watching Steve Jobs’ keynote address and doing some more rooting around on the net, I think I now understand more of what lies behind this surprising decision. 

The problem with IBM's PowerPC roadmap is not so much the lack of processor speed or power, but CPU efficiency. Apple's figures suggest a fourfold difference between power consumption for future PowerPC based processors over Intel's equivalent models. Whilst this is not such a big deal for desktop PCs, such as the iMac or PowerMac (although the liquid cooling featured in the top-of-the-range models do add to their cost), it is critical for portable devices such as laptops and tablet PCs that rely on batteries and are difficult to keep cool without creating additional power drain. Combined with Apple's enthusiasm for dual and quad-core processor machines, this is reason enough for them to get very nervous about the future direction of the PowerPC as IBM have been notoriously slow to deliver thinner, more power-efficient versions of the G5 CPU, a fact that doesn't look set to change in the near future now that most of their clients are using PowerPC variants in non-portable games machines and server hardware (although presumably this could change in the future with the advent of handheld PS3 gaming consoles from Sony).

Conversely, Intel have been steadily reducing their CPU power requirements to the point that the latest Centrino (aka Pentium-M plus on-board wireless chipset) laptops now substantially outperform comparable Apple iBooks and PowerBooks in the battery stakes (the latter are using over-clocked versions of the Motorola G4 processor that negate many of the efficiency gains of the PowerPC design, whilst current G5 designs simply generate too much heat to be used in a laptop at all). Of course performance is still an issue, and a suitable replacement for the AltiVec vector processing unit will need to be found, but it is significant that Apple have not yet announced which Intel processor they will be using. Developer units will be shipping with a 3.6GHz Pentium 4 processor, but as their performance is reportedly inferior to the current generation of PowerMacs, I suspect that we will see some interesting developments in this area over the next 12 months. Apple may even be planning to use a custom built CPU design, or perhaps a version of the Itanium 64-bit processor, which would fit in with their overall strategy of pushing OS X as a true 64-bit computing platform. Conversely, switching to a 32-bit only CPU would be a backwards step that would only be acceptable for the Mac mini and laptops as existing iMac and PowerMac models already support more than 2Gb RAM.

Another odd aspect of this announcement is the timing. Apple made it at their Worldwide Developer Conference on the pretext of notifying developers that they should start porting their products, but at the same time they stressed that in most cases this can be done in a matter of weeks, if not hours or days. Why then are they willing to sacrifice sales of existing product lines for up to a year by announcing that future Macs will be based on a different processor platform? Is it possible that we could see new Intel-based Macs—or ‘Mactels’, as Jobs likes to call them—sooner rather than later? Steve Jobs hinted at this in his keynote address when he stated that new Intel-based Macs would be shipping by June 2006, but that OS X 10.6 (aka Leopard) would not ship until the end of 2006. It looks like the initial wave of Intel Macs will ship with OS X Tiger and could even arrive before the end of this year, although early 2006 seems more likely if, as Jobs says, further PowerPC-based Macs are being released this year (presumably faster versions of the Mac mini and PowerMac ranges?). Another possibility is that Apple already feels that it has sufficient new products in the pipeline to keep revenues buoyant despite a temporary fall-off in Mac sales—iPhone, iPod video, or home media hub anyone? Robert Cringely has even gone so far as to suggest that Apple and Intel may be preparing to merge, although this is a little too speculative, even for me…

The idea that Intel has something other than its processors that Apple is desperate to get its hands on, such as on-chip digital rights management (DRM) support, however, looks increasingly unlikely, although it is almost certainly the case that future Macs will be more than just a bog standard Intel PC with an Apple logo on the front. Aside from the possibility of custom chipsets (which could explain why Apple has gone for Intel rather than AMD processors, which outperform their rivals in many respects), on-board real-time compression and decompression of H.264 (the codec used in QuickTime 7 and next-generation DVD formats) is very likely, and would fit with Apple's plans for a high-definition video and TV delivery platform a la iTunes, but for HD video content. Although I still think it's a pity they chose not to adopt the PowerPC Cell processor that Sony will be using in the new PlayStation 3, there are still plenty of ways that Apple can add value to its hardware without having to devalue its brand. An Intel-based Mac that could run Microsoft Windows as well as offering some unique OS X-only hardware features could give Apple a seriously lucrative product with which to entice potential ‘switchers’, as well as offering Intel a chance to bash their ex-buddies at Microsoft, with whom they are allegedly now pretty pissed (in the American sense).

And this brings me to my final point. Far from being a risky strategy, as many analysts were predicting, Apple's decision to switch to Intel actually puts them in a very strong position. A 64-bit ready Unix-based OS with applications that can run on both PowerPC and Intel hardware gives Apple great flexibility and the ability to shop around for suppliers in a way that was never before possible. Should IBM choose to shift their emphasis and release a low-power version of PowerPC/Cell that wipes the floor with the Intel equivalent, Apple could simply incorporate it into a new PowerMac design. We could even see OS X being released for Sony PlayStation 3 and Microsoft XBox, although I personally doubt that they will ever wish to ship their OS as a purely stand-alone product, as selling hardware is where they make most of their money. However, it creates options, and options are good for Apple. In fact, the only real risk seems to be the effect that last week's announcement looks likely to have on Mac sales between now and whenever the new machines actually appear, but you can bet your bottom euro that they'll be selling like hot cakes when they do finally arrive, especially if they are capable of running standard versions of Windows alongside OS X.

Along with many others, I am slowly coming round to the view that the Apple-Intel deal is a good thing for Apple and a good thing for the future of the Mac. Although it undoubtedly involves some short-term pain, in the long run it gives Apple a chance to substantially increase its market share, and possibly even position OS X as a serious threat to Microsoft Windows, which has dominated the home and business markets so completely for so long. Processors come and processors go, but as Uncle Steve says, the real soul of a Mac is its operating system. I say: The Mac is dead! Long live the Mac! 

Posted on Monday - June 13, 2005 at 02:09 PM            

© 1995-2006