Alistair-Fairweather-640 — Alistair Fairweather

Rumours are circulating that Apple may abandon Intel chips in favour of those designed by ARM Holdings.

Bloomberg reported on 6 November that “people familiar with the company’s research” had said Apple was “exploring” the idea. Of course these sources “asked to remain anonymous”, and Silicon Valley rumours are notoriously unreliable. Yet there is some logic to the idea.

Intel has reigned supreme in the microprocessor chip market since the 1970s. A microprocessor is the brain of your computer — processing the millions of signals required to do everything from editing a document to surfing the Web.

Intel’s chips power hundreds of millions of computers around the world. Together with Microsoft Windows, Intel’s technology has defined the vast majority of the computer industry for four decades.

When Apple moved to using Intel chips in 2005, it hugely increased the appeal of its computers to both end users and software developers. Where software for Macs was once limited to those willing to write for its PowerPC architecture, now software used on Windows computers could be (relatively) easily modified for both markets.

The problem with the PowerPC chips was that they used a different “instruction set” to Intel’s chips. Essentially the two families of chips spoke completely incompatible languages, and thus could not understand (process) each other’s software.

Yet Apple actually moved for a different and more fundamental reason: speed. Motorola, which manufactured Apple’s PowerPC chips, simply couldn’t keep up with the increasing demand for processing speed. Intel, as the market leader in performance, was the obvious choice. The software compatibility was just a nice bonus.

So why on earth would Apple be considering a move away from Intel, particularly after moving chip providers less than a decade ago? Two words sum up the logic: mobile computing.

When Apple launched the iPhone in 2007, it decided against using Intel’s chips. Why? Chips are like cars — when you drive them fast they tend use a lot more fuel and to heat up. Too much heat and they seize and “crash” or even burn out permanently. And without a constant supply of fuel they will quickly grind to a halt.

Intel has long specialised in the muscle-car variety of chips: hugely powerful but also very hot and energy hungry. Both of these things are problematic in a device like a phone, which has a small battery and a limited ability to cool itself down. Computers typically have fans to cool their chips — phones cannot afford that luxury in either space or power usage.

And so Apple chose to use chips designed by ARM, a small but resilient British firm that has been quietly working on its own chip technology since the 1980s. ARM chips aren’t as powerful as Intel’s bruisers, but they use far less power and produce much less heat. They are the Toyota Prius to Intel’s Ford Mustang.

The spectacular success of first the iPhone and then iPad, both of which use ARM’s chips, has prompted some analysts to wonder if Intel’s days at the top might be numbered. This latest rumour seems to confirm that Apple is also thinking that way. Given Apple’s enormous size and influence, just these unconfirmed hints were enough to knock 1% off Intel’s share price and add 2% to ARM’s.

There are some other tempting advantages to Apple moving away from Intel. Apple and ARM essentially collaborate on chip design, giving Apple much more influence over the final result. And by moving its computers to ARM, Apple would have a single chipset across all its devices, allowing for a much more unified customer experience across its platforms.

Apple’s leadership is not scared of making big leaps. Its current CEO, Tim Cook, is the man Steve Jobs put in charge of the switch from PowerPC to Intel in 2005. Even Bill Gates, founder of Microsoft, has admitted he was impressed by how seamlessly Apple managed that difficult transition.

It’s worth remembering that this is all theoretical. ARM chips may be efficient, but Intel chips still outperform them by a large margin. People don’t expect their tablet computer to jump through hoops the way they do their desktop or laptop computer. Tasks such as editing video or large spreadsheets require raw horsepower and Intel is currently well ahead on that front.

And even if you take the rumours as gospel, Apple will only attempt the move in 2017. Presumably its management team is betting that ARM architecture can be scaled up enough by then to service both mobile and traditional computing markets.

It must also be betting that the software available for the ARM chips keeps pace, or it will have a lot of very unhappy customers. Ordinary consumers care as much about which chip their computer uses as which brand of spark plug their car requires. If their favourite software package doesn’t work on the new range of Apple laptops, they will look elsewhere.

And if the rumours are indeed true, Apple will need to replicate its trick from 2005, but with a company roughly 20 times the size and a customer base more than 100 times as large. That may not be quite as simple as in the good old days. — (c) 2012 Mail & Guardian