Microsoft’s original vision for Surface
always included an ARM processor. When the firm unveiled the original Surface RT and Surface Pro, the RT was an ARM-powered device running the then-new Windows 8. Surface RT was supposed to be the low-cost, low-power, tablet option for users who prioritized battery life and wanted a long-lived system with modest performance.
It bombed.
Consumers were confused by the fact that it ran Windows but didn’t actually run any of the applications they expected to run on Windows. Return rates were supposedly sky-high. Microsoft’s Surface 2 continued to use ARM, but the firm moved away from ARM processors with the Surface 3 and stuck with x86 on its other products. Software availability and compatibility appears to be the principle reason why — the Windows Store of the era was a trainwreck and most common applications never surfaced in an ARM-compatible flavor.
The company’s latest Windows-on-ARM effort has been different. Microsoft has already been supporting Windows x86 applications on ARM hardware via emulation, with 32-bit x86 application support, though this isn’t universal. Some applications don’t work under emulation for one reason or another and 64-bit apps are not supported. It’s absolutely possible that the company will unveil further improvements to its underlying emulation engine during the actual Surface event which might end up changing our perceptions of the limits of this device.
Assuming that it doesn’t, it’ll be very interesting to see how well customers respond to this latest effort to bring ARM to Windows. Some publications believe this is an absolutely necessary step. The Verge argues:
The benefits of switching to ARM are manifold. The main one is battery life, which is often rated above 20 hours for a laptop. That’s substantially better than anything Intel currently has to offer. It also makes it significantly easier to add LTE options (and, presumably, 5G) to hardware. ARM processors also tend to run cooler than x86 processors, which frees up manufacturers to experiment with different (read: thinner and lighter) form factors…
So: longer battery life, easier cellular integration, and thinner devices. As nice as the Surface Pro can be, there is a limit to how much it can improve in those areas, and that limit’s name is x86.
The ARM-based Surface does appear to be a thinner machine, as my colleague Ryan Whitwam has discussed, so that angle is covered — but I’m not convinced that this edition of Windows-on-ARM is going to work any better than the last one. Roughly a decade ago, a new type of computer came to market. Known as a netbook, it ran on a low-cost, low-end Atom microprocessor. Lest anyone think I’m remembering the situation via rose-colored glasses, let me be clear: I loathed first-generation netbooks based on chips like the N270. They were bottom-end, ultra-low cost systems in plastic shells and they sold like wildfire thanks to a combination of all-day battery life and what apparently passed for good-enough performance for an awful lot of people.
ARM-based systems, to date, have not replicated this trend. The first wave of Snapdragon 835 and Snapdragon 850-powered systems that came out were praised for great battery life, panned on actual performance, and didn’t sell well. According to Qualcomm, its Snapdragon 8cx platform can compete with an Intel Core i5 in native applications. The question is, how many of those will there be, apart from Office apps? Firefox released a native ARM beta back in April, but a native Chrome version hasn’t been publicly released yet.
It’s hard to predict how people will react in a situation like this. On the one hand, many functions that once required native apps can now be handled in-browser. On the other, people tend to like to know that every application they use on a regular basis will be supported on both platforms.
Back in 2011, I was convinced that Microsoft had to build an ARM version of Windows in order to survive, but the consumer response to Windows RT showed this wasn’t accurate. Offered the chance to buy ARM hardware with Windows, consumers emphatically rejected the idea. They haven’t particularly warmed to it since Snapdragon 850-powered PCs were released. If Microsoft releases a dramatically improved emulator and Qualcomm’s 8cx platform is significantly faster than the Snapdragon 850, we may finally see ARM start chipping away at x86’s market share directly. But there’s another possibility here — one reflected in the actual sales patterns of the hardware people have chosen to buy these past seven years.
Windows PC sales have absolutely fallen, but x86 remains the sole platform of note for Windows. All attempts to split the ecosystem have thus far been unsuccessful. It’s possible that emulation often doesn’t work very well as a market strategy because customers are wary of the idea and find it an unappealing way to preserve app compatibility. This, combined with the inevitable performance hit, may make emulated hardware unattractive.
I can’t speak for others, but I do not enjoy computing across multiple devices and have no interest in owning multiple computers. My preferred number of computers is one. Two is by necessity, not convenience. I am not interested in owning an ARM-based product with less-than-perfect compatibility across the entire realm of x86 computing because I am not interested in being stymied in my ability to run any application I choose on the platform of my choice. These machines would need to hit much lower price points than we’ve seen in the past year in order to be remotely attractive — at least to me. Obviously some people prefer buying the most optimized system for a specific task and might choose to buy this Surface model as well.
Maybe a year from now we’ll all be writing about how this new ARM Surface was the model that finally broke the dam and sparked a wave of ARM-based laptop sales. Maybe it’ll be hailed as the modern version of Asus’ EEE PC when it comes to setting the trend for a new wave of notebook models. Maybe we’ll find out that Microsoft has radically improved emulation, reducing the performance penalty and adding x86-64 support.
But I’d be surprised. I suspect the ARM efforts Microsoft is making will have a marginal impact on the x86 PC market.
Now Read: