Why Windows 10 on ARM Matters (Premium)

Windows 10 on ARM exists because Microsoft has already seen how slowly Intel moves to address real-world industry needs when it has monopoly control of the chipset industry. But Windows 10 on ARM is important­—yes, even to those who don’t see the point of the platform—for many, many reasons. And anyone who cares at all about Windows or Microsoft should be rooting for its success.

That anyone needs to explain this—or to defend Windows 10 on ARM—here in early 2019 is a bit troubling. But based on some unexpected feedback to a tip about an early peek at the native version of Firefox for Windows 10 on ARM, it’s clear that some still don’t understand the need for this platform. And that others are even hostile towards it for some reason.  (And in one weird case, hostile towards me too.)

This is wrong-headed. Windows 10 on ARM matters.

As Microsoft understands, monoculture is bad, and with AMD “circling the drain,” as one executive put it to me, it sought another strong industry partner to play against Intel. It found that partner in Qualcomm, a chip giant that ships far more silicon than even Intel does, albeit primarily on smartphones that are architecturally very different from PCs. Qualcomm isn’t just the biggest provider of system-on-a-chips (SoCs) and other chipsets on mobile, it’s the biggest chipset maker in the world by volume. So it has both the size and technical acumen to take on this work.

Microsoft understands monoculture because it, too, victimized its users, and with them, its partners and others in the ecosystem, when it became too dominant. There are many examples of this, but I’ll just point to Internet Explorer (IE) and the strange, years-long period in which Microsoft did absolutely nothing to advance its web browser because it had achieved market dominance. And because doing so threatened to undermine its then-core platform, Windows. After all, if the web became too sophisticated, users could more easily move to the Mac or even Linux. This is the kind of thinking that occurs in a dominant company, and it is protectionist, not innovative or user-focused.

Intel works the same way in the microprocessor market now that it has achieved the same level of success in PC processors that IE once held in web browsers. With the firm working only to protect this dominance over several years, innovation has come to a virtual standstill because there’s been no scrappy other firm nipping at its heels. And while the PC was never going to push back the smartphone wave, one wonders what might have happened had Intel realized the threat from mobile earlier and had worked to create more efficient designs for tablet PCs and other forms of mobile devices. The past six-seven years of PC decline are equal parts external factors beyond Intel’s control and internal factors that are very much Intel’s fault.

To understand the divide here, you need to understand the needs of the real world. While there are small but lucrative niche markets for very powerful and complex PCs—for gamers, of course, but also developers, videographers, engineers and scientists, and others—there is a much, much bigger market of mainstream users who prefer simpler, more mobile devices. Microsoft, for its part, has made a number of pushes to address this need, including Windows 8 and Windows RT, Windows 10 S and S mode, and now Windows 10 on ARM. But it also needs its partners to make this dream a reality, and Intel—as the primary, almost sole, chipset provider, has been heading in the opposite direction with ever-more powerful chips.

I’ve said and written many times that if Qualcomm and Windows 10 on ARM fail because Intel finally wakes up and starts making more efficient chipsets that offer reasonable performance for the mainstream, enable thin, light, and silent PC designs, and still provide great battery life, then this entire effort was still very much a success. After all, Intel clearly needs this push.

And while I think the future will be a bit more nuanced than that, it’s also pretty clear that Intel has moved only half-heartedly so far to address these needs. The Pentium Gold processor in the Surface Go, for example, is a joke, and that device only gets 4-5 hours of battery life at best. Newer Y-series processors, which have bad reputations for good reasons, are just now starting to make sense, and the battery life is decent too. But some Y-series PCs—like the new MacBook Air—still provide active cooling (i.e. a fan) because the thermals are so tricky. We’re still a generation or two out from this turning into something viable.

Terry Myerson once told me that he wouldn’t rest until those “Intel Inside” stickers were no longer ruining the clean lines of Windows-based PCs. (Microsoft, notably, does not put those stickers on its own Surface PCs.) He didn’t last long enough to see this dream become a reality, but he did at least get the ball rolling on Windows 10 on ARM. Which is Microsoft’s biggest effort yet to end Intel’s PC chipset monopoly.

People criticizing the platform have stated, correctly, that if this is just about better battery life then Windows 10 on ARM was always doomed. After all, some modern Intel-based PCs do deliver stellar, all-day battery life. I’ve reviewed several laptops that provide over 10 hours of real-world battery life in the past year alone, and two recent Y-series entries, the MacBook Air and the HP Spectre Folio, deliver 13 and 17 hours of battery life, respectively. That’s incredible.

But Windows 10 on ARM isn’t just about battery life.

Instead, this platform rests on a combination of factors that together can make for an impressive whole. These factors include, yes, stellar 20+ hour battery life, combined with incredible weeks-long standby times, thin and light form factors enabled by the much smaller physical sizes of ARM-based chipsets and other circuitry, silent and fanless designs, and always-available connectivity via the seamless switching between both cellular and Wi-Fi networks on the fly.

Intel-based PCs can only offer some of these capabilities, but most simply provide beefy U-series chipsets with great performance but in thicker form factors and with annoying fans. And where every single Windows 10 on ARM PC will offer that seamless, always-on connectivity, only a handful on Intel-based offerings do. (In raw numbers, there are actually more Intel-based PC models with always-on capabilities, but that’s due to where we are in evolution the Windows 10 on ARM.)

Windows 10 on ARM and the hardware platform on which it runs are not perfect, of course. Initial systems based on the Snapdragon 835 offered poor, almost laughable performance. And the platform faces massive potential compatibility issues since users cannot run 64-bit Windows desktop applications or install normal drivers and related utilities for hardware peripherals. This latter issue is a “gotcha” moment waiting to happen to any potential Windows 10 on ARM PC buyer.

Qualcomm and Microsoft are trying to address these issues, of course.

On the performance front, the second-generation Snapdragon 850 appears to offer an approximate 30 percent performance boost over its lowly predecessor, and in my testing that puts the chipset right in the “good enough” category, especially if the rest of the PC is properly outfitted (with 8 GB of RAM, for example, and speedy SSD storage.) And the next-generation Snapdragon 8cx, due later this year, will offer a 200 percent performance boost, and will, according to Qualcomm, offer Intel Core i5-level performance. We’ll see, but yes, I am optimistic.

The compatibility picture is a bit more problematic.

Qualcomm is addressing the elephant in the room by ensuring that Windows 10 on ARM users will have access to the most important, and most-used Windows application of all, the web browser. So in addition to Microsoft Edge, which ships with Windows 10 on ARM, it is helping third-parties port both Chromium (the open-source basis for Google Chrome) and Mozilla Firefox to the platform.

But beyond the web browser, backward compatibility with existing desktop applications is still a gray area. While many popular applications—like Microsoft Office, iTunes, and Google Chrome—still ship in 32-bit versions, the Windows world went 64-bit years ago. And many other popular applications—like the Adobe Photoshop Elements example I routinely cite—are only available in 64-bit form. That means they cannot ever run on Windows 10 on ARM. (Which is a 64-bit ARM platform but can only run 32-bit Intel-type desktop or Store apps in emulation.)

This is a problem because these kinds of incompatibilities—be it a popular application or, as bad, a driver or driver utility package—will hit mainstream users the hardest. And Windows 10 on ARM is very specifically targeting that audience. No one using an Intel-type PC ever has to think or even know about whether the PC is compatible with 64-bit apps. Everything just works.

Qualcomm told me they’ve reached out to Adobe and other makers of incompatible 64-bit applications, and they seem to suggest that bringing a 64-bit desktop application to the Microsoft Store in 32-bit Intel (or 64-bit ARM) form is only a bit harder than just recompiling it. But this task is, of course, much more complicated than that. And in many cases may never happen regardless.

Where we stand today is that Intel-based PCs generally offer much better performance and compatibility than Qualcomm-based PCs. But Windows 10 on ARM provides better battery life—often much better battery life—plus generally thinner and lighter and quieter PC designs and killer standby time and seamless always-on connectivity.

The question—as it is so often in such platform comparisons—is which side closes the gap first or at least more effectively. The problem for Qualcomm and Windows 10 on ARM is that performance and compatibility issues can make this platform a non-starter in the eyes of most customers. Few today would—or should—take this risk. So we wonder whether 2019 will be the turning point for Windows 10 on ARM or the beginning of the end.

Intel, like any behemoth, could change the world in dramatic ways if it just woke up and addressed the needs that Microsoft and its other partners have been asking about for years. And it is doing so, slowly. Aside from the aforementioned Y-series chipsets, the hardware giant recently revealed its plans to move to a more efficient 10 nm manufacturing process. (Granted, it’s doing so as the ARM world moves down to 7 nm and even smaller processes.) It could achieve a “good enough” balance and push aside the Qualcomm and Windows 10 on ARM threat.

Again, that outcome isn’t terrible for the industry. But that’s the point: It would never have happened, or would have happened much more slowly, had Microsoft not partnered with Qualcomm with Windows 10 on ARM. This situation could drive more users to alternate platforms like iPad and Chromebook. And that, ultimately, is why Windows 10 on ARM matters. Its very existence will make Windows 10 and the PC better and more viable alternatives to modern challengers. No matter which hardware platform wins.