The complexity of getting the closed binary blobs to run modems and other hardware will make it exceedingly difficult to extract the necessary files and configurations to keep third-party OSes afloat. Then there's the matter of carrier configs, carrier compatibility, expensive carrier certification, and even then, carriers may still just ban the device because they don't like it.
Options will end up being:
- Tearing apart ROMs for blobs and backport/reverse-engineering patches to make them run on alt OSes.
- Find some hardware based on janky Chinese modems that will have little band support, lackluster performance, and likely banned by most carriers.
- Start a new company with the pull to design a new phone OS and hardware with chip and carrier support.
Not impossible, just exceedingly difficult. These systems are heavily integrated and heavily proprietary.
Funny part is, this move will actually make Google lose more money, as Google will lose hardware/software sales, and software dev over this. More people will end up on iOS in the interim, and out of it will come some new mobile OS that will make Google's mobile OS irrelevant in 10 years.
Let's start now, start a company, base a new phone on QNX, have an Android emulation layer for apps until a proper SDK is developed, and just take the wind out of Google sooner than later.
KaiOS is one alternative, it was FirefoxOS. It's pretty sluggish though. Maybe on more decent hardware with some optimizations it'd have a possibility. A lot of Nokia feature phones run it.