That's not really how ActivityPub works though, there's no pulling. They wouldn't be accessing kbin and downloading it's data, it's a push system. We would be pushing copies of our data out to Threads like we do now with all the other ActivityPub services. Threads would then distribute that data to it's users with no extra work on kbin's part. It would just be one more instance in addition to the thousands of instances already out there.

People bring up the XMPP Google situation a lot also, but I think it's a bad analogy for this. Google's adoption of XMPP brought people into the protocol and Google abandoning it took those same people away. Those who were using it before Google could still use it after Google. Anybody who left XMPP to follow Google did it because XMPP failed to adapt to the features people wanted. Thats why we have Matrix now instead.

Wow, had not even thought of that.
RIP to any server that federates with them and has to pay for the bandwidth they're going to need to service Meta's users.