The overwound DNA regions are relaxed by topoisomerase immediately after they unwind to form the replication/transcription bubble, so the entirety of the DNA bar the bubble ends up relaxed, but the underwound region is not completely relaxed. It is constantly relaxed by cutting from topoisomerase as it moves, since underwinding and overwinding have to be met with relaxation (according to Irene) or the process won't be viable energetically or functionally, but it acts as a single point bearing all the energetic strain for the benefit of the rest of the helix, so I doubt it is 100% energetically relaxed. It's definitely being relaxed by topoisomerase the whole time it's moving, though, at both ends.