I would like by Data Lake in Project A to be able to query data in a Cluster in Project B.

Consider an archive scenario when a user of a given app has not logged into the app in [x] number of weeks/months, so all their data is moved to Online Archive. Once they log back into the app again, their "cold" data should now be considered "hot" and be moved back into Atlas. While we can use $out to copy data back to Atlas, there is no current way to remove the "rehydrated" data from S3 once it's been copied back to Atlas

A lot of providers support the same API of AWS. I think it will be simple to integrate them !