The main Carbon contributors are those that ramped down their activities in clang after Google lost the ABI vote, so they decided to create their own toy away from ISO processes.
It's not so much about "the ABI vote", it's the larger direction summarised in P2137 ("Goals and priorities for C++"")
As I understand it, P2137 was written to explicitly spell out the requirements so that there's an actual document saying C++ should prioritise safety and performance over compatibility, which WG21 voted against - making firm the fact that's not what C++ is about.
Left to its own devices, WG21 prefers ambiguity. This is infuriating if you need X, and you tell people "I need X, I can't get that from C++" and they will tell you "No, I'm sure you can have X, maybe the committee just doesn't understand your need" and wasting your time. It needed writing down on paper to ensure there's no room for that ambiguity.
Kate Gregory is one of the key Carbon people and if she had any "activities in clang" I'm not aware of them.
And in Rust one needs to write a RFC, and also not without going through endless discussion threads until a moderator locks them, in some more hot topics.
* In the ISO C++ committee, people lose important votes all the time, again and again. Good features often take years before they gain enough support to be accepted. The committee - despite what the external perception might be - is actually quite hesitant and conservative, at least where it comes to affecting the behavior of existing code. In other words, there isn't such a thing as _the_ ABI vote, it's _an_ ABI vote. So it seems questionable that Google's whole C++ strategy would hinge on that one vote.
