However derived, predictions of global marine species diversity rely on existing real data. All methods, whether based on past rates of species descriptions, on expert opinion, on the fraction of undescribed species in samples collected, or on ratios between taxa in the taxonomic hierarchy, suffer the same limitation. Here we show that infaunal macrofauna (crustaceans and polychaetes) of the lower bathyal depth range are underrepresented among available data and documented results from Australia. The crustacean and polychaete fauna (only partially identified) of the bathyal continental margin of Western Australia comprised 805 species, representing a largely novel and endemic fauna. Overall, 94.6% of crustacean species were undescribed, while 72% of polychaete species were new to the Australian fauna, including all tanaidaceans, amphipods, and cumaceans, as well as most isopods. Most species were rare, and the species accumulation rate showed no sign of reaching an asymptote with increasing area sampled. Similar data are likely for the largely unexplored bathyal regions. This leads us to conclude that the numbers upon which extrapolations to larger areas are based are too low to provide confidence. The Southern Australian and Indo-West Pacific deep-sea regions contribute significantly to global species diversity. These regions and bathyal and abyssal habitats generally are extensive, but are so-far poorly sampled. They appear to be dominated by taxonomically poorly worked and species-rich taxa with limited distributions. The combination of high species richness among infaunal taxa—compared to better known taxa with larger individuals, higher endemism than presently acknowledged because of the presence of cryptic species, the low proportion of described species in these taxa, and the vast extent of unexplored bathyal and abyssal environments—will lead to further accumulation of new species as more and more deep sea regions are explored. It remains to be tested whether ratios of 10 or more undescribed to described species, found in this study for the dominant taxa and for the deep Southern Ocean and the Indo-West Pacific, are replicable in other areas. Our data and similar figures from other remote regions, and the lack of faunal overlap, suggest that Appeltans et al.’s (Current Biology 22:1–14, 2012) estimate that between one-third and two-thirds of the world’s marine fauna is undescribed is low, and that Mora et al.’s (PLoS Biol 9(8):e1001127. doi:10.1371/journal.pbio.1001127, 2011) of 91% is more probable. We conclude that estimates of global species, however made, are based on limited data.