Why Venus lacks plate tectonics remains an unanswered question in terrestrial planet evolution. There is observational evidence for subduction—a requirement for plate tectonics—on Venus, but it is unclear why the features have characteristics of both mantle plumes and subduction zones. One explanation is that mantle plumes trigger subduction. Here we compare laboratory experiments of plume-induced subduction in a colloidal solution of nanoparticles to observations of proposed subduction sites on Venus. The experimental fluids are heated from below to produce upwelling plumes, which in turn produce tensile fractures in the lithosphere-like skin that forms on the upper surface. Plume material upwells through the fractures and spreads above the skin, analogous to volcanic flooding, and leads to bending and eventual subduction of the skin along arcuate segments. The segments are analogous to the semi-circular trenches seen at two proposed sites of plume-triggered subduction at Quetzalpetlatl and Artemis coronae. Other experimental deformation structures and subsurface density variations are also consistent with topography, radar and gravity data for Venus. Scaling analysis suggests that this regime with limited, plume-induced subduction is favoured by a hot lithosphere, such as that found on early Earth or present-day Venus.