Commercially available woven fabrics (e.g., nylon- or PET-based fabrics) possess inherently re-entrant textures in the form of cylindrical yarns and fibers. We analyze the liquid repellency of woven and nanotextured oleophobic fabrics using a nested model with n levels of hierarchy that is constructed from modular units of cylindrical and spherical building blocks. At each level of hierarchy, the density of the topographical features is captured using a dimensionless textural parameter D_n*. For a plain-woven mesh comprised of chemically treated fiber bundles (n = 2), the tight packing of individual fibers in each bundle (D₂* ≈ 1) imposes a geometric constraint on the maximum oleophobicity that can be achieved solely by modifying the surface energy of the coating. For liquid droplets contacting such tightly bundled fabrics with modified surface energies, we show that this model predicts a lower bound on the equilibrium contact angle of θ_E ≈ 57° below which the Cassie–Baxter to Wenzel wetting transition occurs spontaneously, and this is validated experimentally. We demonstrate how the introduction of an additional higher order micro-/nanotexture onto the fibers (n = 3) is necessary to overcome this limit and create more robustly nonwetting fabrics. Finally, we show a simple experimental realization of the enhanced oleophobicity of fabrics by depositing spherical microbeads of poly(methyl methacrylate)/fluorodecyl polyhedral oligomeric silsesquioxane (fluorodecyl POSS) onto the fibers of a commercial woven nylon fabric.