Abstract The organization of convection into relatively long-lived patterns of large spatial scales, like tropical cyclones, is a common feature of Earth’s atmosphere. However, many key aspects of convective aggregation and its relationship with tropical cyclone formation remain elusive. In this work, we simulate highly idealized setups of dry convection, inspired by the Rayleigh–Bénard system, to probe the effects of different thermal boundary conditions on the scale of organization of rotating convection, and on the formation of tropical cyclone–like structures. We find that in domains with sufficiently high aspect ratios, moderately turbulent (), moderately rotating () convection organizes more persistently and at larger scales when thermal boundary conditions constrain heat fluxes rather than temperatures. Furthermore, for some thermal boundary conditions with asymmetric heat fluxes, convection organizes into persistent vortices with the essential properties of mature tropical cyclones: a warm core, high axisymmetry, a strong azimuthal circulation, and substantially larger size than individual buoyant plumes. We argue that flux asymmetry results in a persistent and localized input of buoyancy, which allows spatially aggregated convection to sustain a warm core in a developing large-scale vortex. Crucially, the most intense and axisymmetric cyclone forms for setups where the bottom heat flux is enhanced by the nearby flow and the top boundary is insulating, as long as the convective Rossby number is higher than about 1. Our results demonstrate the great potential for dialogue between classical turbulence research and the study of convective aggregation and tropical cyclones. Significance Statement On Earth, atmospheric convection frequently organizes into large spatial patterns that persist for several days, like tropical cyclones. However, many aspects of this process of organization and its link to tropical cyclone formation are not fully understood. In this work, we use numerical simulations of simple setups of rotating convection without moisture to study the minimal conditions that produce large-scale convective organization, and the spontaneous formation of tropical cyclone–like structures. We find that the latter form more readily for a particular set of controlling parameters and thermal boundary conditions. Our approach seeks to narrow the disciplinary gap between tropical cyclone physics and traditional turbulence research, by bringing together methods, questions, and results that are of potential interest to both.