This is When Windows 10 on ARM Will Finally Make Sense (Premium)

With Microsoft’s new Surface Go a decided no-go on two key metrics, our attention turns once again to Qualcomm and its ARM-powered platform.

I want to root for Surface Go, I really do. I love that Microsoft wants to bring Surface quality to a broader audience. And I’ve never been a fan of the premium-only strategy, where only the most affluent can get the best technology.

But the reality of Surface Go is obvious and negative. The performance is terrible and, worse, the battery life is laughable. Also, its small size relegates Surface Go to children, virtually all of whom would be better off with less expensive and more powerful devices like the iPad.

My sources tell me that Microsoft wanted to use a Qualcomm Snapdragon platform and Windows 10 on ARM for Surface Go. My natural assumption was that the reality of the schedule and the performance and compatibility issues of that platform required Microsoft to look elsewhere. But a trusted source has told me a different story: Intel, it seems, petitioned Microsoft heavily to use its chipsets instead.

In addition to needing to meet a certain price point for Surface Go to make sense, only the Pentium Gold chipset matched Microsoft’s thermal requirements given the form factor. For Surface Go to be this thin and light, and fanless and silent, Pentium Gold was literally its only viable option in the Intel stable.

Look, everything is a compromise. We make compromises every day, in every decision, big or small. But the folks who believe that Surface Go is somehow acceptable in any usage scenario—Brad has said, literally, that it can handle only a single app at a time—are deluding themselves. And I gotta tell you, I’m tired of watching Microsoft’s biggest fans making bad decisions for themselves.

That is, we need to collectively stop pretending that Surface Go makes any sense at all. It does not.

Unfortunately, Windows 10 on ARM doesn’t make any sense at all either, at least not right now. It, too, is a compromise. It’s just the wrong compromise.

In the plus column, Windows 10 on ARM provides epic—literally 20+ hours—of real-world battery life, stunning, weeks-long standby performance, and seamless, game-changing connectivity thanks to its integrated LTE capabilities.

But Windows 10 on ARM suffers from two major minuses that doom the platform in its current state. The performance is abysmal. And compatibility is hugely problematic: It can only run 32-bit desktop applications, leaving major players like Photoshop (even the Elements version) unavailable. And it is incompatible with the millions of hardware device drivers and supporting utility applications that so many rely on.

Some of these issues can be fixed over time: Adobe can, for example, make its Microsoft Store-based version of Photoshop Elements compatible with ARM PCs. Peripheral makers can ship native ARM versions of their drivers and supporting utility applications. And the performance will improve. In part because Microsoft will continue to fine-tune its x86 emulation platform. But in larger part, because Qualcomm will evolve its chipsets over time.

About that.

In the wake of my testing of the Snapdragon 835-based HP Envy x2, which is otherwise an excellent device, I spoke to Qualcomm about the future. More recently, and following an unprecedented Arm roadmap revelation for PC-based chipsets, I spoke with Shrout Research principal analyst Ryan Shrout. So I now have a more concrete understanding of how Windows 10 on ARM will progress over time. And when, finally, it may make sense to actually consider this platform.

And it goes like this.

Someday soon, Qualcomm and its PC maker partners will announce the first Windows 10 on ARM PCs based on the Snapdragon 850 chipset. This is the first Qualcomm chipset to be specifically designed for PCs—the current Snapdragon 835 was originally used in phones—and it will provide a roughly 30 percent performance improvement over today’s Windows 10 on ARM PCs.

30 percent sounds good, on paper. But as Shrout told me, that is a best-case estimate. And it’s not entirely clear how that will translate as a real-world experience. Will the performance improvements be noticeable? Or will it just “feel” faster to the deluded?

Shrout tells me that the performance improvements in the Snapdragon 850 will likely be moderate, not revolutionary, as some still hope. But “a 30 percent improvement is still better than most generational jumps in a processor design,” he noted. And a future generation of chipsets—like the Snapdragon 1000—will offer even more impressive gains.

So why even ship the 850? Shrout theorized that Qualcomm needs to keep the platform humming along in order to satisfy its PC maker partners. If the firm canceled the 850 while promising much better performance in 6 months or longer, many would simply give up on the platform. And they wouldn’t come back.

Plus, the Snapdragon 850 does offer some improvements over its predecessor in addition to that small performance boost, including a 20 percent battery life improvement and even faster LTE speeds. It’s not a complete non-starter.

OK, but what about that rumored Snapdragon 1000? That chipset seems to neatly align with the chips that Arm just revealed in its new roadmap.

Shrout says that PCs based on this design could possibly see a 2X gain in performance, depending on SoC core improvements, process node, and improved app compatibility with Windows. And that kind of massive improvement is what could finally put Windows 10 on ARM—and Qualcomm’s Snapdragon chipsets for PCs—in the right place to be competitive, from a performance perspective, for mainstream consumers.

The question, of course, is timing. It is highly possible that Qualcomm will launch the Snapdragon 1000 chipset in December, and that we’ll learn about the first PCs to use it then or at CES the following January. So it’s reasonable to assume that these PCs will appear in Spring 2019, about one year after the first Windows 10 on ARM PCs appeared.

And that, folks, is when you should finally consider this platform. Qualcomm and Arm will solve the performance problems. And the passage of time will (hopefully) solve, or at least minimize, the compatibility issues. Shifting those two minuses into the plus column nets a huge win, and combined with epic battery life, standby, and LTE connectivity, we have a platform that will finally make sense for many, if not most, users.

I can’t wait. If things work out as expected, things are going to look very different less than a year from now.