I imagine the ATProto infrastructure would survive. Declining activity would be bad for BlueSky, but running a firehose is not that expensive. Many people could fund it themselves. Also, it would get cheaper to run as traffic declines.

Also, the firehose is not strictly necessary either if you’re willing to give up seeing replies from anyone, anywhere and just see the feeds you subscribe to, like an RSS reader.

The hard part is moderation and keeping up with changes in regulations, not the infrastructure itself. Also, continued funding for new development is good. I’m interested in seeing what the permissioned data proposal [1] turns into.

[1] https://github.com/bluesky-social/proposals/tree/main/0016-p...