Abstract The Fe-Ni-Cu sulfide ores at Kambalda, Western Australia, are interpreted to be the result of thermomechanical erosion of underlying rocks by the host komatiite lava flows. However, there is a long-standing argument about the extent of the erosion process, and the degree to which the linear embayments that host the ores were eroded by lava as opposed to formed by tectonic processes. This controversy has fundamental implications for the origin of magmatic sulfide ore, as well as for sinuous rilles on terrestrial planets. The controversy at Kambalda hinges on pinchout features, where sulfide ore at the edges of embayments penetrates laterally into footwall rocks. The most recently published studies of these features interpret them as forming by thrusting of basalts over sulfide-komatiite contacts along the margins of tectonic embayments. Field evidence and X-ray fluorescence element mapping on underground exposures in the Moran deposit demonstrate that sulfide liquid melted its way both downward and laterally into basalt, generating complex plumose melt layers, melt emulsions, and hybridized chromite-decorated contacts. These observations confirm an origin for the pinchouts by thermomechanical erosion, driven by the high temperature, high density, and low viscosity of the sulfide melt. They also provide some intriguing insights into the nature of interactions between sulfide melt and melting silicate rocks in magmatic Ni-Cu-platinum group element sulfide ore deposits in general.