Apple’s announcement of iCloud today reinforces the trend towards infocentricity that I first noted many years ago. And it also shifts the focus away from that core of the computer industry, the microprocessor. Is the computer industry as we know it dead? Yes, it is.
The microprocessor has been at the center of the computer industry since Intel announced the 8086 way back in 1980. Advances in microprocessor architecture and performance have driven computing from that point, but in two different directions. The most important, of course, was the dominance of the Intel architecture continually (and, when you think about it, remarkably) derived from that 8086, and mostly revolving around processor performance – more cores, faster busses, etc. But there’s also a branch that favors low power consumption and low cost, primarily the Atom family in Intel’s case. Intel has a major competitor (AMD) mostly selling the same stuff, and no other architectures matter anymore.
Except one – ARM, which began life as a fairly-low-capability chip for a British PC, again back in the 1980s (1987, to be precise). Intel was once a player in ARM but abandoned this business (selling it to Marvell) in favor of Atom. Big mistake – ARM dominates the mobile device market today, and there’s no end in sight. ARM Holdings uses a business model that is the opposite of Intel’s, doing design work but then licensing to those who build all manner of chips, including the Samsung parts that run in the iPhone and iPad. And the popularity of ARM has led even Microsoft, the other half of Wintel, to port Windows 8 to the ARM architecture. Intel must be very worried here, because there’s a real danger that the volume business won’t go to them – not, anyway, without cannibalizing the cash-cow PC microprocessor business and/or having to really slash prices on Atoms. This isn’t pretty. Back in November of last year, a report in the WSJ had the co-founder of ARM saying that Intel’s microprocessor business is doomed.
OK, maybe a little bluster here, but I’m more than ever in the camp that has computing occurring in the cloud, not on someone’s desk (or, really, in their hand), and increasingly clients will be mobile and thus have a strong emphasis on low-power operation, ARM’s specialty. But, even more importantly, a shift away from running apps on the mobile device to running these in the cloud, which I will continue to argue is essential if for no other reason than that’s where the data is and really should remain for reasons of security and manageability, means that the microprocessor is now effectively commoditized. The ARM business model, I think everyone will agree, works much better than Intel’s in this case. And, sure, Intel will get big chunks of the server business, but the post-PC era of infocentricity is at hand. Lower microprocessor volumes mean that Intel needs to look elsewhere for growth.
And Intel, make no mistake, isn’t going away. We’re just entering the post-PC era, and this is a long-term trend. Intel has plenty of time to respond, and I think they will, eventually going head to head with Qualcomm for handset chipsets – although they are, I must admit, a bit late to the party here.
