The Early Cambrian Kanmantoo Group of southeastern South Australia is a thick marine succession of immature and predominantly turbiditic sandstone and mudstone, and metamorphic equivalents, that was deposited within a continental-margin delta and submarine-fan complex. Deposition was very rapid and occurred just prior to the onset of deformation and metamorphism by the Delamerian Orogeny. The Kanmantoo Group contrasts with older units of the Adelaide Rift Complex (Adelaide Geosyncline) in terms of sedimentary facies, rapidity of deposition, composition and provenance. Importantly, the Kanmantoo Group represents the earliest appearance in Australia of a sedimentary sequence characterised by the 'Pan-Gondwanaland' detrital-zircon signature. Here we present geochemical and Nd-Sr isotope data for sandstone-mudstone pairs collected throughout the Kanmantoo Group. Interestingly, sandstone-mudstone pairs do not have a consistent polarity in terms of their Nd values. This argues against simple mixing/unmixing between two distinct sediment components and suggests a more complex and incompletely mixed multicomponent provenance. There appears to be an abrupt change in isotopic character within the Backstairs Passage Formation in the mid-Kanmantoo Group. Although the origin of this change remains problematic, an overall shift to more negative Nd(†) numbers up-section may reflect progressive exhumation of the provenance region(s) and accessing of increasingly ancient materials. Based on detrital-zircon signature the Kanmantoo Group must be sourced from an area containing abundant 0.7-0.5 Ga magmatism, with other components supplying 'Grenvillian' (1.2-0.9 Ga) and older zircons back to about 3.5 Ga. Large latest Neoproterozoic to Early Paleozoic orogenic and magmatic belts include the Ross Orogen in Antarctica and the Prydz-Leeuwin Belt on the inferred former suture between East Antarctica, India and Western Australia. The former is closer, but lacks known magmatism older than 0.55 Ga. The latter appears to be a more suitable provenance based on the zircon signature, and our Nd data show significant overlap with some existing datasets from along this belt. Recent suggestions that equivalent rocks may extend beneath the central East Antarctic ice sheet (including the Gamburtsev Subglacial Mountains) may provide a more geographically suitable location, particularly from the perspective of paleocurrent data, but are not testable with our present dataset.