Wintertime warm extremes in the high Arctic, primarily associated with intrusions of moist airmasses from the midlatitudes, have occurred with an ostensibly high frequency in recent years. Here, we compute the temperature anomaly return values associated with such events for long return times. Our approach, which we term space‐time maxima–exposure (STM–E), improves on conventional extreme value estimates performed on a single‐location basis by explicitly taking into account the spatial structure of the moisture intrusions driving the temperature extremes. We specifically show that the STM–E approach provides spatially smoother return value estimates, a lower uncertainty in the parameters of the extreme value distribution fits and mitigates the single‐location underestimation of the magnitude of return values of temperature extremes. We compute return values for wintertime warm extremes in the high Arctic, associated with intrusions of moist airmasses from the midlatitudes. Our approach, unlike conventional extreme value theory, takes into account the spatial structure of the moisture intrusions. The figure shows the 100‐year return values for the warm extremes (K), as deviations from the climatology.