Not necessarily. If one can devise a system that will work fully in the open but just be impractical for the competition to implement, it can stay fully in the open while not posing a risk to the original manufacturers' business. For example, if Skype was to move to fully open source, it would still generate revenue for the original Skype inventors because what use is the system when you don't have contracts with telco industries to hook it up to the existing fixed and cell phone lines - sure, one could use their protocol to create a better app and steal from them the user base, but they are not making any money or benefits (except for the brand recognition) from it anyway.

With an OS in question it gets a bit more complicated because OSes are not generally provided as services but rather standalone entities, however I'm sure there is a way to make it profitable enough to justify the efforts of developing it while keeping it open source. There always is a way.

Even if there isn't, if you don't think of yourself as a software but rather hardware provider, it's quite beneficial to you to go the open route and have the thousands upon thousands software developers work for you essentially for free. Why would you care if somebody else can use it if you think you can differentiate enough through the hardware alone?

And in the end, there are quite successful open-source projects that easily reach the complexity and the effort needed for a mobile OS, so how come those are still around and didn't get more closed along the way - projects like Eclipse, the Linux kernel, WebKit, Qt...