Fediverse

They can already access the data, it's all federated and it's all publically available effectively by definition, they don't need to launch a platform that interacts with Fedi in order to scrape it. And Meta will only be able to scrape user profiling data on the people accessing Fediverse through their own tools and platforms. In the large term, all data is useful and getting the additional facets of how their users interact with a twitter-like platform is good - but I don't think that's really why they chose to federate.

But...

What joining Fediverse does offer them is a way of launching their Twitter-rival product with genuine and organic content or activity already present.

Facebook & Instagram's primary demographics are not internet pioneers, they don't tend to build new things - they feed off existing activity and build on top of it. They access the platforms to consume content, and only move to creating or posting content over time as they develop networks on the sites. Meta cannot realistically launch a Twitter competitor whole-cloth. The sort of people who joined Twitter early to build that space aren't joining a Meta product, likewise the people who join new platforms or normal fediverse.

If it launched empty, it would remain empty. People would check it out, see almost no content or no content they care about, and not come back. Meta can only realistically launch a product like Threads with activity already occurring, and things like AI content or fake profiles aren't necessarily convincing enough to lure in the punters. But Fedi is preexisting and active and there's already A Thing there that Meta can point their users at, there's already content to consume and people to interact with.