Well, I thought a bit about this question a year ago, and wrote down the following text. Not that anyone would have been interested at the time, of course.
"Obviously, the open source software itself does not generate any revenues, so the benefits for Microsoft can only be indirect. These benefits might be license revenues for Visual Studio, or ongoing revenues for Azure-based services (which however would require prior investments into NETMF support for Azure access, e.g., AMQP).
You could try to guess market numbers for these additional revenues and build up a business case from that, but I think these numbers would be unconvincing and difficult to defend.
Instead, I would rather argue with “What would it mean if Microsoft gave up support for NETMF?”:
Microsoft would save the cost of a small team (which is peanuts for Microsoft as a company, I dare say).
Microsoft would completely give up its current foothold in the microcontroller space.
Microcontrollers continue to be shipped in higher volumes than the more powerful microprocessors, for years to come. A market forecast by ARM for 2017: they see an opportunity for 4 bn application processors (as used in smartphones etc.) vs. 23 bn microcontrollers, plus 14 bn real-time processors that fall somewhere in between microcontrollers and application processors.
o Thus Microsoft would have no solution for more than half of all the processors shipped by then (note that Windows CE is too large for microcontrollers).
o Can Microsoft, now supposedly a “devices and services” company, afford to be completely absent in this space?
- Many industry observers consider the Internet of Things (IoT) to become one of the most important new trends in IT. The explosion of smart sensors and various types of gateways are only possible due to the inexpensive microcontrollers (and sensors etc., but microcontrollers are among the key enablers of the IoT).
o Without solution for microcontrollers, Microsoft will not have a credible presence in the IoT. Competing platforms will win in this space by default.
o For example, Oracle has started a new push to make Java the language of the IoT, with support from ARM and Freescale. This will allow Java developers to use the same language, library APIs and tools to develop everything from small devices to the cloud. Ironically, this is a major advantage of .NET today thanks to NETMF - (see the graphics at the top of this page for an illustration), but Microsoft would give up this advantage and would allow the competition to catch up and surpass Microsoft.
- One bottleneck for the further development of the IoT is seen in a looming skill shortage, see here for an analysis. Developers who now get acquainted with e.g. a Java IoT platform are unlikely to switch back to a Microsoft offering, if Microsoft should decide to offer something new in a couple of years. Similar to app developers who are already acquainted with iOS or Android: they will not easily switch to Windows Phone.
o Microsoft runs the danger of missing the IoT revolution, like it nearly missed the mobile revolution.
- No one knows exactly which business model will prevail for the IoT. Thus it is crucial to be able to experiment quickly and inexpensively, in a way that allows to later ramp up successful ideas to robust products.
o Here NETMF (and Gadgeteer) can play a useful role, by providing a painless path for experienced .NET developers into the IoT.
o Why not build on this basis, keep a presence in the microcontroller space, and learn from the community about which IoT ideas work and which ones don’t work, and which cloud services or other Microsoft products would help to achieve success?"