The dynamics of the upper thermosphere in the southern polar cap, auroral, and mid‐latitude zones is reviewed. Information has been drawn from published measurements from the DE 2 and AE‐C spacecraft and from ground‐based stations and interpreted using the latest thermospheric general circulation models. Necessarily, frequent reference is made to the F region of the ionosphere, which permeates the upper thermosphere. Great emphasis is placed on the coupling and feedback mechanisms which exist between these interpenetrating fluids. It is shown that to obtain close approximation to observation, a complex model is required which incorporates these mechanisms in a self‐consistent way. Published results from general circulation models show that there is a broad measure of agreement at both high and middle latitudes between simulation and observation over a wide range of geomagnetic activity for the sunlit southern middle and high latitudes near solar maximum. However, there are comparatively few data for a full investigation to be made, and present conclusions must be regarded as strictly interim. Strong driving forces in the auroral zones and polar caps cause a wind regime which is distinctly different from that observed at middle latitudes. Geomagnetic activity has been shown to have a strong effect on the extent of the high‐latitude wind field, its strength, and its penetration to middle latitudes. The orientation of the interplanetary magnetic field has also been shown to be important. The coupling mechanisms between ions and neutrals heat the thermosphere and drive its circulation using energy from magnetospheric sources, but as yet the details of the thermosphere’s variation with geomagnetic and solar activity are not clear from observations. The strong vertical winds reported in the literature are certainly indications of powerful heating, but the connection between observations of vertical winds and the production of gravity wave activity has not been established. Halley, Antarctica, is a critical station in the evaluation of model simulations of the winds in the upper thermosphere since it is a high‐latitude station geographically but middle latitude geomagnetically and is subject to an unusual blend of driving forces. The observations are modeled surprisingly well considering the incomplete accounting of asymmetry caused by the large offset of the geomagnetic and geographic poles, and also of the South Atlantic anomaly. It is probable that, globally, there are detailed disagreements between the actual circulation and the simulations which have not yet come to light for the lack of observations. In the absence of new spacecraft experiments, new ground‐based stations are required in the Antarctic region to enable simultaneous measurements to be made over a wide range of the southern polar cap.