Windows 8 on Intel will be tough to beat
Products based on Intel processors running Windows 8 should continue to claim the lion's share of Windows shipments in 2013 and beyond.
Intel will be as strong as ever despite the emergence of an alternative platform for Windows 8.
That's my forecast after bouncing between Northern and Southern California this week and attending two major tech conferences--the Intel Developer Forum in San Francisco and the BUILD conference in Anaheim.
In the course of three days, I spoke with and listened to plenty of analysts, experts, and industry people. Though the tablet and laptop threat from the ARM camp of Qualcomm, Texas Instruments, and Nvidia is real, I believe Microsoft and ARM hardware companies have a pretty high mountain to climb.
Let's look at the evidence.
Exhibit A: The first Windows 8 tablet. The Samsung tablet that Microsoft distributed at the BUILD conference not ARM. Ironically, most tablets have ARM processors inside because ARM's power envelope is much more suited to thin, small devices.
But that didn't happen, because Windows 8 is not ready for ARM. Yes, Windows 8 will be ready for tablets/laptops using processors from Qualcomm et al next year, but this is the first sign that Windows 8 running on ARM is not going to be easy.
Exhibit B: Compatibility. A corollary of Exhibit A, Microsoft has said the broad compatibility on Intel-AMD won't extend to ARM. So, that obscure (or maybe not so obscure) Windows application that a business needs may not run on ARM. In other words, Windows as we know it (broad backward compatibility) will be there, as always, for Intel/AMD but not for ARM. At least not at the beginning.
In fact the best argument made to me was by a Wall Street analyst (speaking on background so I can't give his name). When businesses buy a Windows 8 tablet, they will probably buy Wintel (Windows-Intel): that's because they'll get assured compatibility and security. Most businesses will not take a chance on an ARM-based tablet, however inexpensive and power efficient.
And businesses may not even opt for a tablet. Buy late 2012, early 2013 (when Windows 8 is expected) Intel-based Ultrabooks may be close enough to a tablet experience (who knows what an Ultrabook will look like in 2013) that the tablet, as we know it today, may not be as attractive--or necessary.
This argument, however, doesn't apply to consumers. The price for an Ultrabook will have to come down to below $800 (dare I say $700?) to compete with a Windows 8 ARM laptop.
Exhibit C: Performance. Intel is all about performance. An Ultrabook based on thein 2013 will be an extremely attractive option compared with a Windows 8 tablet from, let's say, Acer running on a Qualcomm chip. That's not to say that an ARM-based tablet or laptop will be slow, but Intel will have the upper hand in performance. And count on Intel (and AMD) having all-day battery life to boot. (It bears repeating, though, that if there is a big price gap between an inexpensive Windows 8 ARM laptop and a pricier Windows 8 Ultrabook, performance may not matter.)
Which brings us to an Intel weakness: Atom. Let's face it, that power-efficient Intel chip design does not seem destined to compete very effectively in the tablet space. At least not in the next 12 months. And this was made very clear to me at BUILD. I asked a person from Samsung why it opted to go with a Core i5 chip--not Atom--for the Windows 8 tablet. To paraphrase, this person said Atom was way too slow and simply not an option.
Moreover, Intel needs to make sure that 3G/4G is part of its chip offerings going forward. Qualcomm integrates 3G into its chips, and Nvidia is headed in that direction. To offer an Ultrabook without 3G/4G in 2013 will be like offering a laptop today without Wi-Fi, i.e., a deal breaker. Intel's purchase ofshould be able to make this happen.
Finally, below is a BUILD conference interview with Bill Crean, an OMAP product marketing manager at Texas Instruments (OMAP is TI's ARM design). Right next to him was a glass case showing off TI's Windows 8 tablet. Toward the end of the interview (at the 2:10 mark) he explains why he believes quad-core ARM processors aren't necessary yet and that operating systems don't (yet) take advantage of the extra cores. This is in response to Nvidia, which has been touting its quad-core Tegra 3 processor for tablets and high-end smartphones.