I have to disagree with the answer above based on my own observations. Just to make sure I wasn't hallucinating, I just tried a file copy across an NT network and the full file size (542,702,345) appeared on the target server as soon as the file header was read as the transfer was starting. However, the space available on the target drive was not updated until the transfer was complete.
I don't believe the space is being pre-allocated because if you run out of space, the transfer doesn't crash until you actually reach the point in the transfer where there is no more space on the target drive. It seems to me if it were pre-allocated, file transfers would always crash at the beginning rather than some point in the middle or near the end.
I believe that what is happening is that the initial directory listing on the target drive is based on the information in the file header, subject to correction when the transfer is complete. While the directory entry is there from the start, the entire file is not until the transfer actually completes the copy of the last byte and the file allocation table has been updated with the list of clusters that make up the file.
paul