What would Apple do with an ARM-based SoC?

Apple is up to something, and as usual it’s the skills and press releases of other companies that shine the best light on Apple’s new direction.

Predicting the future is never easy, and with Apple it’s a special challenge. Apple always delivers far more, or far less, than people expect. They are experts at evolutionary change, like the new Macbooks. However, they are also experts at revolutionary change. The iPhone stands as a shining example.

Recently, Apple’s moves in acquiring PA Semi bought them a great deal of expertise in embedded systems, low power processing, and more generally in the area of SoC semiconductor design. As recently discussed, Apple possesses a lot of ARM design capability for a company that isn’t an ARM licensee. So, it is a reasonable assumption that they are.

The reasons to believe this are compelling. With the upcoming ARM Cortex processors in dual and quad-core configurations, ARM is going to be pushing out of the top end of their traditional performance envelope, and challenging Intel’s Atom for the netbook performance crown. More interestingly, ARM already thrashes the Atom on MIPS/Watt performance, and we know this is so important to Apple that Steve Jobs spent several minutes discussing it during his 2005 WWDC keynote – it was their primary reason for switching to Intel processors.

Now we have a new datapoint: Apple has taken a financial stake in Imagination Technologies. This is interesting for many reasons. Imagination has expertise in high performance graphics processing on small power budgets – something Apple desperately needs if they intend to bring high performance graphics to an ARM device for netbook or embedded use. The iPhone is passable, but it doesn’t have the performance to be a mainstay game platform, to do much with 1080p HD, or make for a pleasant OS X desktop experience.

Imagination Technologies is also a member of the Khronos Group, yes, the same people who recently finalized the OpenCL specification that is a big part of Snow Leopard.

From this we can start to draw reasonably safe conclusions about the device Apple would be designing, and its capabilities would betray its possible usage in future Apple products. What can we conclude?

Apple seems to be designing a high performance, power efficient, ARM-based SoC. It likely has high performance graphics, and an exceptional power envelope. It likely has the performance to support OS X natively, with comparable performance to current dual core Intel-based laptops.

This device seems almost overpowered for iPhone use. It would be a nice fit for a proposed netbook, where it would provide good performance, long battery life (days not hours), and an amazingly small thermal and physical footprint. It would also be very comfortable in the AppleTV, where it could reduce manufacturing cost and product complexity, and increase the AppleTV’s capabilities.

The greatest new product potential I see, when looking at the direction Apple has taken their core business over the last few years, is in selling devices that are tied to revenue streams that Apple controls. I’m referring, of course, to iTunes and the App store.

An AppleTV and a Wii have basically identical hardware requirements, and their functions support each other amazingly well. This new, proposed device would enable such a system to be built for less than $200. If you’ve used an iPhone, an AppleTV, a Wii and a DVR, you could see how these things coming together would be a Good Thing™.

The buzz, though, is about an Apple netbook. What does this proposed chip allow? If we take both design paths, the options are a giant iPhone with a touch-screen or proper keyboard, or the thinnest, lightest and ‘funnest netbook ever!’

What do you think?