I think that will depend largely on how closely the order of the inserted data matches the order of the clustered index. If it matches then SQL will be writing sequentially so there's no concern about there not being space where it wants to put the new row. If it doesn't match then there's the possibility that there won't be space where it needs to put it and fragmenting will happen. One thing that I don't have time to right now but would be interesting to is seeing as this is all happening in one insert whether or not SQL will order the rows in the order it needs to write them if no sort order is applied.