Will Intel (INTC) ever dominate mobile in the same way it now dominates PCs and servers? That's the question investors in Intel have been asking themselves over the last few years. Ever since ex-CEO of Intel Paul Otellini turned down Steve Jobs, when he asked Intel whether it wanted to supply Apple (AAPL) a chip for the iPhone and then again for the iPad (now that's how you miss mobile!) Intel has been frozen out of mobile almost entirely. So what exactly is the problem? Why can't Intel replicate the success it has had in PCs and servers in tablets and smart phones?
Intel's last big Apple win
Intel wants to win Apple's mobile chip business, and then some. But how did Intel win Apple's last big piece of business? The answer to that question should help illustrate why Apple is struggling now in mobile.
Currently one of the big obstacles for Intel winning Apple's iPad or iPhone chip designs is that these devices run on ARM Holding's (ARMH) chips. For Apple to switch to Intel would require it to either run its existing iOS apps on a x86 powered emulator, which would come at the cost of a performance penalty, or require all of Apple's app developers to recompile their apps for the x86 instruction set language.
Now back in 2005, Apple did decide to switch to using Intel's chips in the Mac instead of PowerPC chips. PowerPC was an alliance by IBM (IBM), Motorola and Nokia (NOK), and the chips used RISC architecture, just like ARM Holding's chips do today. So why was Apple willing to change back then and not now?
Mobile's double bind
There are a number of reasons why. First of all, IBM simply couldn't deliver the chips that it had promised Apple it would design for them. In 2003 Steve Jobs had announced that Apple would have a 3GHz G5 processor by mid-2004. This was the time when gigahertz was the be all and end all of performance. By the start of 2005 IBM still hadn't delivered anything. In addition:
- Back in 2005 it was much easier for Apple to change horses so to speak because the Mac OS X developer community was much smaller and more loyal to Apple than is the case with iOS. In fact Mac OS X wasn't even compatible with the Mac's predecessor operating system, OS 9. So Mac application developers were used to being put upon.
- The typical Mac user then was far more sophisticated about computing than the average iPhone or iPad user is now. This meant that they could be expected to know the workarounds to all the bugs and issues that the transition to Intel chips entailed.
- Finally, and this is the crucial issue, PCs (and servers) can force software limitations by simply using more powerful hardware. So if a Mac user found the emulation program, called Rosetta, was running his programs too slowly for his liking, he could upgrade to a more powerful machine, at some cost of course. The problem with mobile is that it faces a double constraint: more powerful chips are more expensive, but they are also more power hungry, and good battery life is a key feature for most people.
Moreover, Intel can't count on the ARM "alliance" to be anywhere near as weak as the PowerPC alliance. The PowerPC alliance back then was rather like AMD is now: starved of R&D investment. And this was wholly down to the fact that by 2005 hardly anybody was buying PowerPC chips.
Intel could have been VHS
If Paul Otellini had accepted Apple's terms for supplying the iPhone's chips in 2007, history would have been very different: iOS would have been rewritten for the x86 instruction set architecture, most likely exclusively for x86. Arm would have lost out on a huge amount of business and consequently would've have been a much less formidable competitor going forward. Microsoft (MSFT) would have no reason to create Windows RT or have Windows Phone run on ARM chips. And Google (GOOG) may have decided to optimize Android for x86 rather than accept the performance penalty of running both on ARM and x86 chips.
Now all is not lost for Intel's mobile hopes. Intel can compete on a level playing field with ARM chips in Android devices. The reason it has failed to compete here is the result of other factors. In phones specifically, it has had trouble matching Qualcomm's (QCOM) cellular baseband processor chips, and in tablets its chips have so far lacked the required integrated features (Wi-Fi, blue-tooth, etc) for the devices it has been trying to win.
Thus Intel's problem competing with ARM based designs in Android devices is down to an unfavorable competitive position rather than a question of industry standards. Essentially, Intel has struggled because its process technology advantage has been trumped by either better or more relevant chip design. And that's not because Intel's chip designers aren't any good. It's simply because there are many more companies in the ARM camp ready to supply chip designs to the verisimilitude of mobile devices than Intel can handle.