Instead it got kicked out because Kent constantly ignored the kernel's contribution rules and is unlikely it will ever be accepted back into the kernel.
And it went in when it did because Redhat was pushing for it and claiming to be supportive - but that never materialized. They wanted to get something for free without investing, or putting in the absolute bare minimum.
A _lot_ of people were saying publicly and privately "dear god yes we need something better than btrfs" - but no one from the existing kernel community was interested in stepping up.
Community's still growing, though. A lot of people have gotten active in making sure bcachefs actually works well for people end to end, and there's a hell of a lot more to shipping a filesystem than just writing kernel code.
The FS was marked experimental, so there is no urgency in fixing bugs or providing features in a certain cycle. Everyone using it knows what they got themselves into. You can still provide the DKMS module for faster fixes and features for anyone who wants to use BCacheFS more seriously for the time that the upstreaming process takes, but eventually it would have all been on mainline.
Asahi is taking a similar approach where they have their downstream kernel and push things upstream once they are mature.
That means the upstream kernel is not useful for running on that hardware now, but things are moving there eventually.
All this has been discussed to death, we don't need people armchair quarterbacking a year later. It's over, it's time to move on.