Could? Should? Would? – Mobile OS

The topic of mobile operating systems (OS) has been largely settled for over ten years. The world is split between Apple’s iOS and Google’s Android. And while there are a few others out there, the market has been in stasis for a long time. Could that change? Should it? Would any of the stakeholders want it to? There are some reasons to think it might, and we have a purely speculative suggestion for how things could go.

Building an operating system is expensive. Google reportedly employees 1,577 people on the Android team. and while we cannot vouch for that source, that sounds about right. You would need 500 engineers just to build the code for the kernel and another 1,000 to build all the other pieces and provide limited support for all the hardware requirements, really limited support (that is still our top viewed post many years on). All in, it probably costs something like half a billion dollars a year to develop and maintain a mobile OS. And then it does not generate any revenue directly. Microsoft is really the last company to charge people for an OS directly, and even there it is bundled into subscriptions and other services. Apple monetizes its OS through hardware sales and Google through ads and data, a model which is now under many regulators’ scrutiny.

Apple is never going to give up control of iOS, that runs counter to pretty much everything the company stands for. Google, on the other hand, seems decidedly ambivalent about Android. In 2006, when they acquired Android, controlling a mobile OS was of existential importance. The mobile web could have gone in a very different direction. However, the world today is very different. Google need Android to make sure its core assets – search, maps, Gmail, etc., the “G-suite” – had a position in mobile. This is now firmly entrenched, and Android is only a small part of that position. Google is by far the leader in mobile search because they pay Apple a few billion dollars a year to maintain their position as default search on iOS. And keeping the other handset makers onboard is as much about offering the G-suite as it is about Android itself. With the regulators increasingly questioning Google’s bundling of Android to G-suite, it might make sense for Google to spin off Android and go to a direct payment model to the handset makers instead. Admittedly $500 million to maintain Android, versus billions to buy search default is not great math, but Google could afford it and this would likely solve other problems along the way.

Google’s ambivalence is on display elsewhere. The most recent Google Pixel phones, which are supposed to highlight the best, “purest” version of Android have been hugely underwhelming, lagging the market in terms of feature and performance by a considerable margin. Google, and Apple, also seem to be radically changing the digital ad ecosystem in ways that may further call into question the importance of Android.

Maybe, and this is a big maybe, it is time for a new approach. The easiest path would be for Google to truly open source Android. Spin it off to someone like the Linux Foundation. Google would lose the ability to mandate device search, but as we suggested above there are other ways to accomplish that. The question then becomes who will push Android forward. True, there are a lot of people in the community who would want to contribute code, and the Linux Foundation has a good track record of building coding communities, but this is complex stuff and will require some deep-pocketed sponsors.

Fortunately, there is a large group of companies who care a lot about mobile operating systems. First and foremost are the chipset makers, especially Qualcomm and Mediatek. One of the great mysteries of mobile is why Qualcomm never did this. They tried for many years prior to the advent of smartphones to popularize BREW, a mini-OS for feature phones. The early success of first iOS and then Android came so quickly that the market was settled before Qualcomm could formulate their response. And while this was a great missed opportunity, the market probably would have blocked them one way or another back then. That being said, times have changed.

One of the unspoken forces driving open source software has been the support of other parties in the value chain. Linux is the best example. Linux was for many years heavily subsidized by Intel, in the form of cash as well as engineers on Intel’s payroll who spent their time contributing code. The success of Linux opened up the server market which shifted value to the semiconductors that powered those servers, semis which Intel built into its massive data center group, source of half their profits today.

There are clear parallels with the mobile OS today, but also differences. For one, Qualcomm does not have a lot of competition anymore, at least not at the high end of phones. When Intel first started supporting Linux the server market was much more fragmented, cloud computing and commodity hardware were not even ideas. The mobile world has largely standardized on Android. Secondly, it is not clear that open source software is a viable way to build a consumer-facing OS. As much as Linux was heralded as an alternative to Microsoft Windows, it was really an alternative for Microsoft’s server OS and tools, and to this day Linux desktop environments are very much a niche market.

But Qualcomm is not the only one with an interest. The handset vendors are all either too small to afford their own OS or lack the ability to build reliable software but would love to have more control, any control, over the software that powers their phones. .

Add to this mix the brewing dispute between Apple and China’s leading Internet platforms – especially Alibaba and Tencent – who seem to be playing some form of chicken over ‘privacy’, among other issues. Those companies have a big interest in the mobile web, obvious software capabilities and deep experience with open source projects. Add all of these together and there is a viable path for achieving an open sourced Android.

Will it ever happen? Only if Google leads the way. The industry is too fractured for any one party to start from scratch. Qualcomm could propose it, but Mediatek would oppose it. Same goes for Ali and Tencent. But if Google had a good reason to open source Android, it could turn out to be for everyone’s better.

Leave a Reply