Intel's CEO on stage when Apple switched to Intel's X86 CPU architecture
Website semiaccurate has been circulating the rumor that Apple was going to dump Intel completely from its line of computers, laptops and desktop, when chips based on 64-bit ARM processors designs will be ready for mass production. This seems like a wild guess, but the site, known for being sometime accurate insists that this will happen.
This raises a number of questions: Is it possible and what would that mean to Apple users? Why would Apple do it? What does Apple need to do to make this successful? Finally, what would that mean for Intel? Let’s try answering these questions.
Is it possible and what would that mean to Apple users?
Yes, it’s technically possible. At any given point in time, Apple can decide to move to a different processor architecture. It has done so in the past when it went from PowerPC to Intel’s X86. However, this was not an easy change. Apple had to “emulate” the Power PC processor (this slows things down) and support both platforms for a number of years. Eventually, the company stopped supporting PowerPC completely, but the process wasn’t painless and the advantages of going with Intel were obvious: that allowed Apple to use the economies of scale from the PC industry… against itself. Great move.
Limited app library: current applications won’t run on ARM without an emulator (slow) or a recompilation (which takes a long time for software vendors to build, test). In short, the new ARM-based Apple computers would have a much more limited usage model at first.
Slower performance: it’s extremely unlikely that ARM can compete with Intel in terms of raw performance in the next 5 years, so the rumor is based on the idea that performance is “good enough”. The argument is not new… but it has been made for the past 20 years or so. Some time ago, it had been estimated that a Pentium II at 600Mhz “fulfilled all the needs” that most users have. Yet, users have kept buying ever more powerful computers since, and still want computers to be faster.
Longer battery life: yet, the ARM architecture has been historically able to deliver lower performance-per-watt ratio, and that means that if you can make due with a slower computer, your battery life could be extended significantly. Note that at some point, the display becomes the power-hungriest element in the system.
Slimmer, lighter laptops: ARM-devices also tend to be smaller and lighter because they require less cooling. Most are fanless and that reduces noise, thickness and some power consumption.
Use iOS and Mac OS in the same chassis: Having iOS in Apple laptops would be a killer app because you get a true Instant-On device that is as productive as a computer, with a relatively light OS and apps. Of course, you would still need a classic OS like Mac OS or Windows next to it for app compatibility, but having the ability to turn things on and off is also great to save battery life. Today, Apple can’t easily use iOS without working on major software/driver modifications, or without adding an iPad touch to its laptops. In both cases, it costs money and creates problems that Apple probably doesn’t want today.
A move from Intel’s X86 to ARM could bring some tangible benefits, but it would also bring many headaches.
Why would Apple do it?
While the user benefits is Apple’s ultimate goal, the company will want to align its own interest with any radical change in platform. So, how could an ARM architecture benefit Apple?
Money: Apple currently buys chips from Intel, and although they probably negotiate the pricing, Intel remains their sole supplier of X86 chips… Of course, Apple could also reach out to AMD and do some classic vendor balancing. However, regardless of who sells the chip, someone is going to make a decent margin on them.
If Apple was to design, produce and use its own ARM-based chip, it won’t have to pay a markup to itself. That’s the classic benefit of vertical integration. However, there are also some downsides: Apple will have to deal with the chip manufacturing (outsource it to a foundry, manage it), and most importantly, it won’t benefit from Intel’s advanced manufacturing process, which everybody agrees, is the best in the world.
Differentiation: with total control over its hardware platform, Apple can accelerate its differentiation against the competition and add features that it deems necessary.
What does Apple need to do to make this successful?
The problem with switching to a new architecture is that you may need to deal with its legacy, forever. Look at Intel: the main added-value of X86 is that you can run any software that was built for any X86, ever. The latest X86 processor is backwards-compatible at the binary level with all its predecessors. That’s a *huge* advantage, and one that no-one else has been able to provide. It’s also a moot advantage in a mobile world where there is no legacy apps. But this will soon change.
While Apple does not “have to” abide by such stringent standards, users would like want it to. However, it is likely that if Apple was to break compatibility every once in a while, they can probably get away with it.
To make such a transition successful, Apple needs to bring tremendous added-value to make users swallow the pill. It’s got to have “good” performance and *extraordinary* battery life and design. If you take the Macbook Air as an example, it is a level of performance that can be achieved by upcoming ARM architectures, so it’s a legitimate possibility.
What that would mean for Intel?
While it is a wild rumor, it is always interesting to have a “what if” discussion. If Apple was to switch to ARM, what does it mean for Intel? The answer is : it depends.
If Apple is the only player to move, well, life isn’t so bad for Intel. However, with Windows 8 coming for ARM processors, a successful move from Apple can incite other OEMs to reciprocate that strategic move. Over a long period of time, this could be a real problem in terms of growth prospects.
Of course, this is “if” there was a very successful transition from Apple…
Conclusion
While the rumor is an interesting topic of discussion, I don’t believe that Apple will move in a massive way to ARM and “dump” Intel. The “massive” or “dump” part is just a dumb notion. The smart thing to do is to introduce one iconic product, maybe a new Macbook Air, that will transcend the benefits of having an ARM architecture. Most likely, those benefits will be in the industrial design (thin, light, beautiful) and in the battery life (10/15 hrs of *real use*).
By slowly introducing a new architecture, Apple would have time to study the benefits and user acceptance, and let developers come up with new version of their apps.
The move isn’t without risks for Apple. For one, Intel has been continuously making strides in low-power processors, even if ARM designs are currently firmly ahead. From my point of view, Intel can go in ARM’s territory more easily that ARM can go in Intel’s territory. Windows 8 will provide some real insight into this standoff. In my opinion, Intel has been “artificially” slowing the evolution of Atom by fear of eroding its (extraordinary) margins. Yet, Atom still has better margins that most semiconductor products.
Secondly other ARM silicon vendors can beat Apple at this game, and Apple could find itself outgunned: a situation that the company knows very well from its PowerPC days. Although it can get by with its software (Mac OS/iOS) and design differentiation, the company has learned that “specs” do matter as well.
Yes Apple can do it, if it wants to. However, any transition won’t be as “dramatic” as some believe (or have you believe).
Related: SoC the force that rules all smartphones and tablets
Links: semiaccurate, ARM,