The mechanisms that result in the formation of high-pressure (HP) and ultrahigh-pressure (UHP) rocks are controversial. The usual interpretation assumes that pressure is close to lithostatic, petrological pressure estimates can be transferred to depth, and (U)HP rocks have been exhumed from great depth. An alternative explanation is that pressure can be larger than lithostatic, particularly in continental collision zones, and (U)HP rocks could thus have formed at shallower depths. To better understand the mechanical feasibility of these hypotheses, we performed thermomechanical numerical simulations of a typical subduction and collision scenario. If the subducting crust is laterally homogeneous and has small effective friction angles (and is thus weak), we reproduce earlier findings that <20% deviation of lithostatic pressure occurs within a subduction channel. However, many orogenies involve rocks that are dry and strong, and the crust is mechanically heterogeneous. If these factors are taken into account, simulations show that pressures can be significantly larger than lithostatic within nappe-size, mechanically strong crustal units, or within a strong lower crust, as a result of tectonic deformation. Systematic simulations show that these effects are most pronounced at the base of the crust (at ∼40 km), where pressures can reach 2–3 GPa (therefore within the coesite stability field) for millions of years. These pressures are often released rapidly during ongoing deformation. Relating metamorphic pressure estimates to depth might thus be problematic in mechanically heterogeneous crustal rock units that appear to have been exhumed in an ultrafast manner.