ABSTRACT Freely available nucleic acid sequencing databases have accumulated to a vast archive of genetic diversity, in excess of 50 petabase-pairs from tens of millions of experiments. Together, these data constitute a digital survey of Earth’s genome. However, the richness of biological information contained within these repositories remains largely unexplored, in large part owing to the technical challenges of analyzing petabytes of data. Recently, Logan completed the sequence-assembly and compression of 27 million sequencing libraries from the Sequence Read Archive (SRA), and here, we systematically search Logan-SRA to reveal the global diversity of the DNA-based Papillomaviruses (PVs). In a single ∼10-hour alignment-based search against the Logan assemblage, we independently re-identified 65% of the 992 PVs recorded within the NCBI Virus database, a body of work representing over five decades of PV characterization. We further expand the diversity of PVs by 34%, identifying 383 novel PV types spanning 105 associated host species, including taxa with no previously associated PVs, such as rhinoceros, voles, and grey foxes. Through integration of virus phylogeny, sample geography, and ecological metadata, we show that novel PV discovery is not directly proportional to sampling effort, and there are significant hotspots of PV biodiversity in East Africa and South America, and that undersampled biomes can yield disproportionately more novel PVs. Public sequencing repositories contain vast, unrealized biological information that is now accessible through advanced computational infrastructure. Here we lay the foundations for the efficient analysis of petabase-scale datasets for DNA virus discovery, with applications to uncovering all genetic diversity.