Android Things is a Failure (Premium)

Google announced this week that it would limit Android Things to smart speakers and smart displays only, effectively killing any chance that it will turn into a general-purpose Internet of Things (IoT) platform.

“Given the successes we have seen with our partners in smart speakers and smart displays, we are refocusing Android Things as a platform for OEM partners to build devices in those categories moving forward,” the Google developer advocate Dave Smith revealed via the Android Developers Blog. “Therefore, support for production System on Modules (SoMs) based on NXP, Qualcomm, and MediaTek hardware will not be made available through the public developer platform at this time.”

Google originally had far grander plans for Android Things, which was announced in December 2016 as a “comprehensive way to build IoT products with the power of Android.” The idea at the time was that developers familiar with Android app development could use their existing skills and knowledge to expand into this new market for connected devices.

But Android Things was late to market and has always been poorly supported. At the time of its announcement, even Microsoft’s IoT platform, since undercut by Azure Sphere, had broader hardware support. But then over a year and a half passed. And by the time the first non-beta version appeared in May 2018, Microsoft’s two IoT platforms still maintained distinct advantages over a system that, frankly, should have been a no-brainer for developers given the popularity of Android.

Key among Android Things' limitations, of course, was its short three-year support time frame and its lack of a comprehensive security service like that Microsoft offers for Azure Sphere.

Google, of course, has seen success in one key area that has eluded Microsoft and most other IoT platform makers: Smart speakers. So, it makes some sense, I guess, that this would be the focus going forward. (Smart displays are really just smart speakers with a display; as such, you might see them as the same basic type of device.)

Google says it will continue to support “popular hardware like the NXP i.MX7D and Raspberry Pi 3B” so that developers can “experiment” with building smart, connected devices. And it still allows makers and other non-professionals to target up to 100 devices for free. Those with more scalable needs can look to Google’s Cloud IoT Core platform for secure device connectivity. And Google is, of course, addressing Azure Sphere, in this case with a coming Cloud IoT Edge suite of managed edge computing services.

That’s not super-interesting to me or, I suspect, to most Thurrott.com readers. What is interesting, however, is this situation in which a dominant platform maker builds a new platform, based on the dominant platform, that subsequently fails. And here, I think, the parallels between Microsoft and Google are obvious.

Microsoft dominated the personal computing market when PCs were the so...