Why is Microsoft neglecting the Windows on Arm ecosystem?

Windows on Arm was Microsoft’s hope for a post-Intel future, with devices that sip power but offer complete connectivity. With Microsoft’s next Surface event looming, is the platform still relevant today?

Microsoft expected developers to leap at the chance to compile their applications for its platform after the first Windows on Arm machines emerged in 2018. After all, as Apple has since ably demonstrated, there are benefits to moving away from traditional Intel-based architecture. 

We noted improvements in performance and power efficiency on macOS machines running on Apple silicon, while the desktop experience was unchanged. Could the same be true for Windows on Arm?

In short, no. Not with Microsoft’s relative neglect when compared to its x64-based stablemate. The firm has squandered the opportunity to move users on from an Intel-based norm and has dithered in the process.

While, for better or worse, Apple forced developers to migrate to its M-series CPUs, the majority of Windows applications have continued to target Intel and AMD due to simple market forces and the absence of clear guidance from Microsoft.

Microsoft’s Windows on Arm efforts have moved in stops and starts. Its exotic – and expensive – Surface Pro X has been gradually updated over the years, culminating in a rebrand in the form of the SQ3-powered Surface Pro 9 in 2022. Microsoft's partners have also had a go with their takes on the architecture, with varying degrees of limited success. However, an inexpensive option – a real alternative to the x64 platform – for the masses has yet to materialize. 

Then there was the attempt to lure developers with the promise of Project Volterra, unveiled at Build 2022, and finally released as the Windows Dev Kit 2023. Since then, things have gone rather quiet. I have a Windows Dev Kit 2023, as it happens, and the ownership experience hasn’t been a good one.

Everything that’s wrong with Windows on Arm

The Windows Dev Kit 2023 comprises a Snapdragon 8cx Gen 3 along with 32GB RAM and a 512GB SSD. On paper, it is an impressive bit of kit. It even bears a distinct resemblance to Apple’s Mac Mini, except at a keener price point and an external power brick.

But, after almost a year of usage, I’m sad to report that if the Windows on Arm dream isn’t dead, then the Windows Dev Kit 2023 doesn’t offer any promising signs of health.

We’ll start with the hardware. On my machine, some things simply don’t work properly. The clock, for example, runs slow and requires synchronization. Getting output from the Mini Display port connector is a little random; I’ve had to sacrifice a USB-C connector in the name of reliability. Sometimes it wakes up, sometimes it doesn’t – and I need to hold down the power button sometimes.

These are all niggles one expects from a first-generation device, and renaming Volterra to the ‘Windows Dev Kit 2023’ implies Microsoft knows it isn’t ready for the market.

One would’ve at least expected some of the issues to have been ironed out in the months since release, though, and yet they persist. Even as Windows itself is updated, the Windows Dev Kit 2023 encapsulates all that’s wrong with the Windows on Arm program.

Then there is the software. It took until November 2022 for Microsoft to release a fully supported version of Visual Studio for Arm64 – a release that came after the Windows Dev Kit 2023 became available.

Even users keen enough to try Windows Insider builds on the device are frequently disappointed when Microsoft opts not to offer a build to Windows Insiders with Arm64 hardware with little or no explanation. Hardly a developer-friendly approach.

The disappointment continues when using the device. It doesn’t offer a snappy experience when compared with the Mac Mini, which has less RAM and storage. While it doesn’t quite feel like wading through treacle, as with previous Windows on Arm devices, it’s certainly sluggish.

To make matters worse, running Windows on Arm in a virtual machine (VM) on Mac – we used Parallels – is a better experience than using Microsoft’s own hardware. All of this leads us to believe that, perhaps like its consumers, Microsoft is losing faith in Windows on Arm. While Microsoft has dithered with the operating system, Intel and AMD’s chips have improved. Battery life is no longer as catastrophic for Intel-based devices compared to Windows on Arm hardware, eroding the advantages of Arm-based chips.

Behind all of this is the specter of Arm in the Data Center, arguably a much better place for it in the Microsoft ecosystem. After all, with more apps delivered via the browser, is there much point in having Windows developers produce native Arm code, if not to run on a server?

Microsoft has a final opportunity in this year’s Surface event to make good on its Windows on Arm promise. Devices need to be inexpensive and reliable, and the company must demonstrate Windows on Arm is a viable option for users rather than adventurous power users. If it doesn’t, one can only conclude the company has little interest in the platform.

Windows on Arm: Unfulfilled potential

Ignoring Windows RT, a short-lived and locked-down version of Windows for Arm chips that turned up in early Surface tablets, Windows on Arm (or Windows on Snapdragon) finally arrived in 2018 in the form of Lenovo’s Yoga C630 WOS.

The device enjoyed superb battery life, 4G connectivity, and the look and feel of a proper Windows device. But it had some issues, and it didn’t take long for customers and reviewers alike to realize. The Qualcomm Snapdragon 850 processor might have been able to cope with lighter productivity tasks, but users attempting serious work soon found themselves yearning for an Intel or AMD-based alternative. 

The biggest headache was one of emulation. As soon as a user attempted to move outside of the Arm app ecosystem, the experience of Intel emulation (assuming it worked) was more akin to swimming through honey. Sure, it was an improvement on the total absence of x86 support from Windows RT, but users were still presented with an experience some way off that available from similarly priced Intel or AMD-based kit.

Even more astonishingly, large swathes of Microsoft’s own software had not been recompiled to run natively. Its Edge browser required emulation. As did Teams. It would take until 2020 before either emerged in native form.

Microsoft tried again in 2019 with the Surface Pro X – despite lacking important native software – and hoped to bring some of the Surface magic to the beleaguered platform. One intention of the Surface line was to show manufacturers how Microsoft reckoned its software could be packaged. In that regard, it has largely succeeded.

Could the same be applied to the Windows on Arm ecosystem? With the Surface Pro X, Microsoft set out to find out. But while Arm devices were traditionally associated with cheaper hardware, the Surface brand was not. Furthermore, Microsoft decided to ask top dollar for the Surface Pro X, powered by the company’s own take on Qualcomm’s chips in the form of the SQ line. The problem was you could grab a high-end Surface Pro 7 for a similar price, and it was arguably a good deal more useable. 

The platform has continued to languish as the years have rolled by, with Microsoft occasionally tossing it a bone in the form of a native version of Visual Studio, or other productivity apps that really should have been ready at launch. 

It’s difficult to see Microsoft's approach as anything other than neglect; if it doesn’t seem interested in the technology, why should its customers be?