The Passing of Wintel? (Premium)

I’m a long-time fan of Jean-Louis Gassée, the former Apple exec and founder of Be, Inc. But his most recent missive needs to be addressed.

For the record, Gassée has been blogging about personal technology since blogging was a thing. I recall reading his thought back when Be was still an ongoing concern. And I’ve always been fascinated by his take on the industry.

These days, Gassée blogs at Medium, and as a paid subscriber, he’s one of the few voices I actually follow regularly on the service. He’s still got it, and he’s as interesting as he ever was.

(For whatever it’s worth, I still think Be would have been a better fit as a next-generation MacOS than was NeXT, though getting Steve Jobs back into the company obviously paid, rendering that argument moot.)

Yesterday, Gassée published a controversial post, Apple Silicon: The Passing of Wintel, to Medium in which he argues that the Mac’s transition to Apple Silicon will “force” Microsoft to “polish” Windows 10 on ARM, a move that will in turn cause PC makers to “reconsider their allegiance to x86 silicon and have serious consequences for the old Wintel partnership.”

Many will likely have an immediate and visceral reaction to this notion, and I will admit that I saw the post yesterday, read through it quickly, and then bookmarked it so I could go over it again today. Not surprisingly, his post was covered broadly today, especially in the Apple-centric blogs that would lap up such a thing unthinkingly.

But since I’m currently reviewing a modern Windows 10 on ARM laptop, I feel particularly well-positioned to consider what Gassée is proposing. Because today, Windows 10 on ARM is nowhere near ready for primetime despite years of improvements to the hardware and the software.

And while I appreciate Gassée’s premise, he’s wrong about the choices that Microsoft—really, the PC industry—has. He claims that Microsoft can either “either forget Windows [10] on ARM and cede modern PCs to Apple, or forge ahead, fix app compatibility problems and offer an ARM-based alternative to Apple’s new Macs.”

But those are not Microsoft’s only choices, and in this case, I sort of wish that Gassée had read me. As I’ve noted in the past, Windows 10 on ARM exists for one reason and one reason only: To goad Intel into making more efficient mobile chipsets. Microsoft doesn’t care if Qualcomm ARM wins. It just wants Intel to have some competition so that it will show up and do its best work. (And in this arena, Microsoft’s experience rings true: Intel, like Microsoft, doesn’t do well when there’s no viable competition.)

The good news? Intel does have competition now. It’s just that it’s not coming from ARM, at least not in the PC space, it’s coming from AMD. And AMD, which already led the way to x86-64 (the 64-bit version of x86 that it calls x64), forcing Intel to follow, could do it again.

Of course, the problem for Microsoft and Intel—and AMD, really—is that the volume part of the personal computing revolution is in smartphones, and smartphones are powered by ARM. So even if we get modern mobile CPUs on PCs, the PC is never going to be more than just a sliver of the broader market. Going ARM won’t change that. In fact, the cost, pain, and time required to make such a transition speaks to the problem with even trying. Why bother?

I do think that Intel and AMD will continue to take ideas from ARM—like the big/little core architecture—just they previously took ideas from RISC and incorporated them into their CISC-based chips for PCs. And that in doing so, we’ll see more efficient mobile PCs. But will Intel/AMD actually license the ARM architecture and create its own ARM-based chipsets for PCs? Again. Why bother?

But here’s the thing Gassée really gets wrong: He refers to the Intel world as being high margin, which implies that ARM is low-margin. It’s not: The pricing of middling ARM-based PCs is stratospherically high, as you can see from the Flex 5G I’m reviewing: The ARM version starts at $1400, compared to just $500 for the Intel version and $670 for the AMD version. And this isn’t just a problem in the PC space: Qualcomm’s chips are so expensive that Google isn’t even going to use flagship chipsets this year so it can save money on its new Pixel handsets.

The point is that Intel and AMD already provide low-margin offerings for PC makers, and Qualcomm can’t match that pricing. That the Qualcomm chipsets are slower and don’t have amazing battery life benefits in addition to their compatibility problems explains why Windows 10 on ARM has failed.

Look, Apple Silicon is just a different animal, and because Apple makes both the hardware and the software, it can lower its own costs by switching the Mac to this platform. But in the PC space, that’s not actually necessary since neither Microsoft nor any of the PC makers own both the hardware and the software. The competition in this market, both from silicon providers and from PC makers, is what keeps prices low.

In other words, Gassée is comparing Apples to oranges. Or at least to PCs. And what works for Apple isn’t necessarily—isn’t almost ever—the solution its competitors can adopt.