Doesn’t that just fracture the audience/ users?
Networking, telecoms, and the Internet all tend to erase space and topology. Relatonships of federation, following, muting, and blocking re-impose a topology. It's not necessarily geographic, but it exists and it profoundly affects information flows through the network.
Put another way, on both Mastodon and Twitter, there is the ability to block or mute other profiles. This creates interesting dynamics about where information does and does not flow, and who has control over that.
In the case of Mastodon, there's additional refinement to that control in that individual instances can set several levels of resistance or friction between other peer instances. Posts might simply be muted on the public timeline, or an instance might be entirely blocked. What levels of moderation are enacted and which instances are affected varies by instance. One consequence is that an instance administrator bears consequences of their own actions regarding moderation and permitted / tolerated behaviour, as other instances may determine their own interactions on account of such behaviour.
In practice, what I've noticed across several social networks, having a blocking mechanism, is that the old saw holds true: friends come and go, blocks accumulate.
In the case of Google+, a network I both participated in and studied reasonably extensively, one of the greatest predictors for whether or not an account was closed or banned was how active it was. The higher the level of activity, to a large extent, the greater the odds of it later not existing. This was based on a few snapshots I'd made and/or explorations of high-profile accounts. I suspect mechanism tended both to be burnout by some high-profile users and tendency for such usage to violate site norms (spam, abuse). It was interesting to see though.
As long as no one company, person, or government owns and operates the internet, it will always be "fractured".