
The Windows community is understandably excited by Snapdragon X and Copilot+ PC. But let’s not get ahead of ourselves. This platform isn’t mature or full-featured enough to completely replace Intel and its x86/x64 processor architecture. And while the microprocessor giant has been struggling lately, it’s still dominant in PCs and has already weathered the potent threat it’s faced with Apple Silicon-based Macs. Arm may be the future of the PC, but that’s only one possible outcome. Indeed, it’s far more likely that Intel will simply sweep Arm aside as it improves x86/x64 yet again, this time to be more efficient.
The benefits of Arm are undeniable in general. But the specifics make this architectural shift problematic with Windows. Today, as ever, Windows 11 on Arm is a subset of Windows 11 (x64). Most Windows apps run in emulation, and no x86/x64 Windows drivers work in this new system. This will not change.
I’ve written this before, but it’s an important point: When Apple shifted from Intel to Apple Silicon (Arm), it left the old architecture (Intel) behind. It did this previously with its Power PC too Intel shift, and the Rosetta-based emulation layer that enabled the shift was eventually retired. But that’s not what Microsoft is doing. It’s maintaining two forks of the Windows 11 code base, one for x64 (as I’ll now call it) and one for Arm64. Both are identical for the purposes of this discussion—there are obviously low-level differences—and both will be identically supported, with the same features and functionality, going forward … forever. Well, for the foreseeable future. Nothing is permanent.
We can debate the complexity of this side-by-side development, but it doesn’t matter: This is just a market reality. Today, all Windows PCs out in the world are running x64 from a statistical perspective. Tomorrow, this December, a year from now, that will almost certainly be true, too: Even if Snapdragon X Copilot+ PCs are somehow the sales sensation of the year, even the giddiest forecast puts their market share in the very low single digits. Windows 11 on Arm isn’t becoming the mainstream Windows version anytime soon, if ever.
To understand the challenge, let’s first look at Apple Silicon. Many likely believe that the M-series Macs are the inspiration for this shift to Arm, but if you’ve been around long enough, you know that’s not the case: Microsoft shipped its first Arm-based version of Windows in 2012, with Windows RT. It restarted these efforts with Qualcomm and Windows 10 on Arm in 2017, and the two have spent the past several years improving the hardware and software behind the platform.
It’s not even fair to describe Apple Silicon as a wake-up moment for Microsoft: It had been calling on its silicon partners—Intel, especially, but also AMD—to rethink their approach to microprocessors for years and embrace the instant-on, reliable, and efficient mobile paradigm that had subsumed personal computing starting with the iPhone in 2007. For Microsoft, Apple Silicon was the ammo it needed to convince Intel to finally wake the f#$k up. And it worked: Intel has been working on more efficient new hybrid x64 architectures ever since.
But forget that history: Apple Silicon exists. Apple has shipped three generations of M-series chip families for the Mac, and has expanded its offering to include higher-end chips for workstation-class products like the MacBook Pro. A fourth generation, the cunningly named M4, is curiously limited to, and wasted in, the latest iPad Pros. But everyone knows that Apple is plotting an AI/NPU-based resurgence, and the M4 family will almost certainly start appearing in Macs as soon as this year.
But how has Apple Silicon impacted the Mac? Has it resulted in massive gains in market share?
Here are the numbers. Apple released its first M-series Macs in late 2020. That year, Apple’s Mac controlled 7.9 percent of the PC market by unit sales (market share), up from 6.8 the year before. In 2021, the Mac’s market share was still 7.9 percent, so there was no growth year-over-year. In 2022, the Mac lost market share, falling to 7.2 percent. And last year, the Mac accounted for 8.7 percent of the PC market. In four years, Apple Silicon triggered a 0.8 percentage point gain in Mac market share. That’s it.
(If you’re a glass half full type, you could argue that the Mac saw a double-digit unit sales gain of 17.2 percent during this time. Fair enough. But as always, I will point out that it’s easy to grow share when you’re tiny.)
Whatever you’re thinking, Apple Silicon has not triggered a wave of Mac switchers.
This is good for Intel and AMD, and it’s good for PC makers, which have long been more concerned with fighting each other than taking on Apple. Apple has had its influence on the PC market: PC makers have long flocked to premium PC designs with higher margins as a way to help the bottom line, and all major PC makers now offer various models to counter the MacBook to some degree. But PC makers have been using this strategy for years. This is why you see so many high-margin/low-volume PCs, like workstations, gaming PCs, and now PCs aimed at so-called creators. The power of the upsell is real.
What Apple offers, of course, is that elusive power-per-watt efficiency benefit. Which is as important as it is hard to understand and market to mainstream users. It’s so hard to understand that most in my part of the industry don’t understand it either: The Verge published an article just yesterday that purports to compare “every Snapdragon X chip against the Intel Core Ultra, AMD Ryzen 8000, and Apple M3,” but it completely ignores its central point, “a new focus on power efficiency.” This article does not even bother to describe the relative efficiency of the Snapdragon X compared to its competitors. There are tables with specifications, lots of them. But nowhere does it address efficiency, be it performance per watt, power draw on idle vs. load, power draw on battery vs. plugged in, whatever. This publication has no idea what it’s talking about.
But that’s fine. No one does. What people really want to know is simple. How’s the performance? How’s the battery life? How’s the compatibility? And how loud is that damn fan?
These are, of course, the topics I’ve been exploring since I got my first Snapdragon X Copilot+ PC about 11 days ago. And my experiences so far, while overwhelmingly positive, betray a few hard facts.
The performance is largely excellent, one might say comparable, when pitted against mainstream PCs based on Intel Core/Core Ultra chips.
The battery life is better than those PCs, but not as good as we hoped and not as good as Apple delivers with its latest generation MacBook Air. Snapdragon X-based laptops deliver 10 hours or less of real-world battery life, where the Air delivers fully 50 percent more battery life, or about 15 hours.
Software compatibility is good, certainly better than previous Windows 11 on Arm efforts, but with a major caveat: Everyone reading this will have at least one important app that will not work, and that means you’ll have to switch apps or not use Snapdragon X. There’s also a smaller caveat: Most apps are not native to Arm64, so they run under emulation. This is good, it works well for the most part. But emulated x64 apps impact battery life. (Here, I must be vague: I don’t know by how much.)
Hardware compatibility is also good, but here again there is the same caveat: Many will have hardware peripherals that do not work on Windows 11 on Arm, either well or at all. And here, there is no out. If it doesn’t work, it doesn’t work. There’s no workaround.
The fan noise has not been an issue, at all, on any of the three Snapdragon X-based PCs I’ve used. And for whatever it’s worth, these PCs are all, to a one, far thinner and lighter than Apple’s behemoth MacBook Pros, each of which also has fans and fan noise under load. These Snapdragon X PCs are always compared to the MacBook Air, even by Microsoft and Qualcomm, but they often compete effectively against the MacBook Pro too. Depending on your needs and workloads, of course.
So that’s mostly good news. For fans of Windows 11 on Arm like me, it’s breathtakingly good news. But what this really means is that we finally have a set of PCs that delivers most of what makes the MacBook Air special, including one quality, reliability/consistency, that I did not highlight above because I don’t think most people even think that way. And that’s great for those people who used to gravitate towards the now moribund Ultrabook class of laptops. The thin, light designs that business travelers favored before COVID-19 upended everything.
The thing is, these types of computers are less necessary today. Many work from home now or split time between home and an office. Many who used to travel extensively for work, like me, no longer do so. We’re usually near a power outlet, and most PCs support fast charging. So what we’re really looking for collectively—yes, there are outliers—is some combination of uptime and performance. It’s nice not worrying about battery life, for sure. But it’s also nice knowing you can plug in at any time too.
Anyway, it’s competitive. It’s in the ballpark, as I say. The MacBook Air, in particular, still has important advantages over Snapdragon X PCs related to battery life, much less emulation, and fewer compatibility concerns. There are some unknowns around reliability and more. But overall, it’s where we need it.
But Apple Silicon didn’t really change the PC industry. So what exactly does Snapdragon X give most of the market, the people and businesses that were never swayed by the Mac in the first place? Aside from addressing the Apple envy stuff, what really changed here?
It depends.
For those users who prefer, thin, light, and ideally silent portable PC experiences, Snapdragon X is really neat. It’s a bit Wild West right now because of lingering compatibility issues that may or may not ever be solved. And it could improve over time as more apps are (hopefully) ported to Arm64. It’s done a nice job of slotting into a space where Intel and AMD largely fall short.
But Snapdragon X doesn’t scale like Intel’s and AMD’s x64 designs. Its CPU and NPU are, for the most part, top-rated, but its GPU is unimpressive, and there are no dedicated GPUs on mobile, let alone desktop systems with add-on cards, now or in the foreseeable future. (Apple faces this same issue.) So Snapdragon X can’t replace many—or most—of the PCs that people are now using. It’s not just uncertain, which is scary for mainstream users. It’s just not possible in some cases. Most obviously with gaming PCs, which require powerful GPUs and full compatibility with all games, but in workstations and any other high-end market. For creators, it depends. Some creator workloads work well on Snapdragon X, but some do not.
Snapdragon X is like the rookie baseball player that hits a home run the first time at bat. That’s exciting, and fun. But there’s a whole career coming, and it’s too soon to suddenly claim that they are better than the savvy veteran sitting at the end of the bench. They may never be that good. It’s too soon to say.
And yet, some are claiming that for Snapdragon X already. That it’s game over for Intel, especially, if it doesn’t get its act together. As if Intel hasn’t actually been working on doing just that for the past few years. Hasn’t, in fact, been experimenting so chaotically with ever more aggressive hybrid designs that it’s temporarily harmed its chips’ reliability, performance, and efficiency in its mad bid to catch up. Intel looks broken to us on the outside, I see it too. But it is naïve and premature to count them out.
I will just point to two important truths before moving on.
First, research Intel Lunar Lake and see what the company is doing to dramatically shift how its PC microprocessors work. And then ask yourself: If a Lunar Lake-based PC can work as well—performance, efficiency, reliability—or better than a Snapdragon X PC but offer full compatibility with decades of apps and hardware devices, then what exactly is the point of Snapdragon X again?
And back to that efficiency thing: I noted that no one seems to understand what this even means, and I pointed to a Verge article that claimed to address it but does not. Everyone reading this likely assumes that Snapdragon X-based PCs are super-efficient because of the Arm platform, and that these PCs must somehow be sipping or expending power in some minimal way. But that’s not true: From a power consumption/usage perspective, Snapdragon X is roughly as inefficient as the current generation Intel Core Ultra chips (it varies based on model and workload). And it is less efficient than the MacBook Air M3: Where that chip consumes just 20 watts under a full load, the Snapdragon X hits 80 watts or more depending on the model. That’s Intel Core Ultra 9 H-series territory. Yikes.
Here, again, I’ll just point out that it’s in the ballpark. And that we should be happy with that. But let’s not exaggerate the gains here. Snapdragon X and Windows 11 on Arm are a huge advance when compared to their predecessors. But there is so much work to do, and so many future Intel, AMD, and Apple Silicon chip designs coming that seek to undermine these gains. These are established players with the engineering know-how to make it happen. And it’s on them to deliver. But we cannot discount the threat.
I celebrate what Microsoft and Qualcomm have accomplished. And I acknowledge that we’re in a slice of time, a bubble, and that things will change. Nothing is certain.
Claiming otherwise just makes you look silly.
With technology shaping our everyday lives, how could we not dig deeper?
Thurrott Premium delivers an honest and thorough perspective about the technologies we use and rely on everyday. Discover deeper content as a Premium member.