In theory, yes. In practice, you probably don’t want to shave that particular yak. Since there isn’t a magical “take this from here and put it there” API, you would still have to:

1. Download all the appropriate blob data from the source repo
2. Upload all the blob data to the destination repo by creating individual blob objects
3. Create the recursive forest of tree objects necessary to reference all the blobs you just uploaded
4. Then create a commit object to root the trees you just planted on the appropriate branch

Or, you could have Git do it because it knows how to manipulate its own database.