Dwarf planets including transneptunian objects (TNO) and Ceres are >500 km large and display a spheroidal shape. These protoplanets are left over from the formation of the solar System about 4.6 billion years ago and their study could improve our knowledge of the early solar system. They could be formed in-situ or migrated to their current positions as a consequence of large-scale solar system dynamical evolution. Quantifying their internal composition would bring constraints on their accretion environment and migration history. That information may be inferred from studying their global shapes from stellar occultations or thermal infrared imaging. Here we model the equilibrium shapes of isolated dwarf planets under the assumption of hydrostatic equilibrium that forms the basis for interpreting shape data in terms of interior structure. Deviations from hydrostaticity can shed light on the thermal and geophysical history of the bodies. The dwarf planets are generally fast rotators spinning in few hours, so their shape modeling requires numerically integration with Clairaut's equations of rotational equilibrium expanded up to third order in a small parameter m, the geodetic parameter, to reach an accuracy better than a few kilometers depending on the spin velocity and mean density. We also show that the difference between a 500-km radius homogeneous model described by a MacLaurin ellipsoid and a stratified model assuming silicate and ice layers can reach several kilometers in the long and short axes, which could be measurable. This type of modeling will be instrumental in assessing hydrostaticity and thus detecting large non-hydrostatic contributions in the observed shapes.