"Theorem A" in this paper by Y. Gordon: http://www.math.uiuc.edu/~mjunge/Comps/Gordonm.pdf is a comparison inequality for Gaussian processes. Is there an analogue of this result for subgaussian processes?