The End of Microsoft’s Android Fever Dream (Premium)

In the heady early days of Windows 10, Terry Myerson’s team did what they could to reverse the decades-long decline in native Windows app development. It would prove to be a futile, losing effort that first focused on righting the wrongs of the mobile app platform his predecessor had introduced in Windows 8. But the next step was trying to convince developers to bring their codebases from other platforms—Android, iOS, Win32/Windows desktop, and the web—via a series of so-called bridges, each of which confusingly worked quite differently.

The Android bridge, code-named Astoria, was perhaps the most interesting because it included a way to run Android apps directly in Windows 10 without requiring the developer to port their code. This effort was killed by Microsoft because—get this—it was too good, and Myerson’s team knew that it would undermine their new native app platform, which had been renamed to the Universal Windows Platform (UWP).

In the wake of this internal reshuffling, Microsoft acquired Xamarin to gain access to its Xamarin Forms solution, which allowed developers to write apps for Windows, Android, and iOS using a single codebase. Xamarin Forms eventually morphed into .NET MAUI, and it remains Microsoft’s only native cross-platform client app solution for developers. But by the time that happened, Myerson was long gone, the bridges were all dead, and Microsoft had given up trying to push UWP on developers. It migrated unique UWP capabilities to the desktop as “Project Reunion,” later renamed to the Windows App SDK, but developers have been ignoring it ever since, just as they did with UWP and its predecessors. As noted previously, these efforts were all futile: In an era defined by web and mobile apps, new native Windows app development hasn’t been a viable option for developers for decades.

Finally facing this reality, Microsoft and the Windows team recalibrated: Since native Windows app development no longer mattered, developers were free to use other operating systems and often did. Those who focused on iPhone and iPad apps would naturally choose a Mac, since that was the only truly viable option (cross-platform solutions like Flutter, MAUI, and React Native notwithstanding). And most other developers chose Linux. It is perhaps ironic that Microsoft’s Visual Studio Code contributed to this shift by being available across platforms.

But there was a new goal to make Windows the most desirable platform for developers. This effort is ongoing today, and it reached the heights of absurdity when Microsoft made Dev Home, a developer dashboard of sorts, part of the base installation of Windows 11 along with related features like Dev Drive. But before we got to that extreme, Microsoft introduced, in turn, its Windows Subsystem for Linux to help keep developers who might have otherwise switched to that open source platform in Windows, and then the related Windows Subsystem for Android for Windows 11. Though there was a developer story there, WSA was positioned as being more consumer-focused than WSL, with a promised “store within a store” experience that would help users run Android apps side-by-side on Windows, as is the case in Chrome OS.

But there was, as always, a problem: WSA only worked with the woeful Amazon Appstore for Android and its lackluster subset of apps, and not the Google Play Store that might have given this functionality a chance. But with Google and Microsoft continually at each other’s throats, that was never going to happen, and the WSA soldiered forward, as unused as most of Microsoft’s client advances of the past decade or more.

Until now, that is: Today, Microsoft announced that it will kill WSA with support officially ending in one year, on March 5, 2025.

“The Amazon Appstore on Windows and all applications and games dependent on WSA will no longer be supported beginning March 5, 2025,” a note on the Microsoft Learn website reads. “Until then, technical support will remain available to customers. Customers that have installed the Amazon Appstore or Android apps prior to March 5, 2024, will continue to have access to those apps through the deprecation date of March 5, 2025.”

I could be wrong—faux outrage runs strong in our community—but I can’t imagine that many will mourn WSA’s passing. It was lackluster from the start, with few decent apps and an incredible collection of what can only be called crap. Even Google’s Play Games offering—which lets Windows users run a selection of Android games on Windows—is of more interest. And you’ve probably never even heard of that. (It is not coincidental that Google has been very clear that Play Store Games is in no way associated with WSA.)

So what are we to make of all this?

Microsoft’s various attempts to make Windows relevant as a native apps platform over the past 20 years have all failed: The Windows Presentation Foundation (WPF, called Avalon briefly during the early days of Longhorn), Metro (Windows 8), and UWP were all non-starters with developers, and the best we can hope for today are classic desktop apps with modern WinUI 3-based user experiences built with the Windows App SDK. Microsoft’s more recent attempts to make Windows more relevant to developers are still an open question, and while I feel that the company has done all it can, it’s clear that developers will use what makes the most sense and that, in many cases, Linux or the Mac meet those needs.

WSA is a failure that ties into both of those efforts and one might wonder now whether this technology, in some ways, is a symptom of the issues with each. Apple likes to brag about the virtuous cycle that benefits its intertwined platforms, and Google experiences some of that as well between Android on phones, tablets, and other devices and Chrome OS. But Microsoft does not benefit from this virtuous cycle, and even those developers who prefer Windows see little in the way of platform symbiosis.

No, I think this is about AI.

For years, Terry Myerson’s Windows team was obsessed with finding what it called “the next wave,” that platform that it could monopolize and dominate as it had during the PC era. Microsoft had “missed” mobile, which was a misnomer given that Apple had in fact reinvented mobile with the iPhone, kicking off the era that displaced Windows and the PC as the primary personal computing device. And Microsoft had, in turn, fumbled its efforts in subsequent waves-to-be like Mixed Reality and personal digital assistants.

You don’t have to be particularly sophisticated to see that Microsoft is now betting the company on AI. And that, using the language of the Myerson era, AI is indeed the next wave. This is where all the innovation is happening, both in the cloud and on the client, and Microsoft’s developer efforts, as we’ll see in full force at Build 2024 in May, will be entirely focused on AI. Not mostly focused. Entirely focused.

There will still be language about making Windows the best choice for developers. But in truth, Microsoft doesn’t care what platform developers use as long as they target its AI, both on the client and in the cloud. It is not coincidental that its AI tools for developers focus largely on the web and on its cross-platform Visual Studio Code. Its consumer- and business-facing AI tools likewise work in Windows or mobile where it makes sense—primarily via its dominant Office applications—but mostly on the universally accessible web. The answer to “Where do you want to go today?” is no longer Windows, it’s AI.

There was no AI story for WSA, so it had to go. And I will bet good money that WSA is only the start of a great reshuffling in Windows, one that’s tied to Microsoft’s new focus. I suspect this might mark the end of Microsoft’s Android device efforts as well. You’re either with AI, or you’re out. And WSA is out.