Each device has it's own firmware and software
Let's dissect this for a moment: Each device is accessed and controlled via Google Home, regardless of whether it's Chromecast Audio, or Generation 1, 2, or 3 Video, the Ultra, or
Chromecast built-in. As for the firmware, they all run some variant of Chrome OS.
Given my experience as a consulting software engineer, I feel quite confident that the firmware simply disables various codecs based upon the state of some hardware register, which would inform it as the whether the device actually supports audio/video or just audio. That's exactly how I'd code it if I were in charge of the project, as it would make everyone's job much easier only having to maintain one code base across various dongles.
Based upon all the above, I'm quite confident that the protocol, and the various dongles that utilize it, will be usable for years, if not decades, to come...