Robert Miller, global director of books for the internet archive, stated in a documentary in 2013 that there had been an estimated 100 million books published in the world, that the archive had an initial target of 10 million, and that their book warehouse had space for 3 million. Based on those figures, 500k is a rather large number. Maybe some of those 500k are duplicate scans?

[1] https://archive.org/details/archive_documentary_internet_arc...