When a liquid drops impinges a hydrophobic rough surface it can either bounce off the surface (fakir droplets) or be impaled and strongly stuck on it (Wenzel droplets). The analysis of drop impact and quasi static "loading" experiments on model microfabricated surfaces allows to clearly identify the forces hindering the impalement transitions. A simple semi-quantitative model is proposed to account for the observed relation between the surface topography and the robustness of fakir non-wetting states. Motivated by potential applications in microfluidics and in the fabrication of self cleaning surfaces, we finally propose some guidelines to design robust superhydrophobic surfaces.Some plants leaves and insects shells exhibit extreme hydrophobicity, making the deposition of water drops on their surface almost impossible [1]. All these superhydrophobic biosurfaces share two common features: they are made of (or covered by) hydrophobic materials, and are structured at the micron and sub-micron scales.During the last decade much effort has been devoted to design artificial solid surfaces with comparable water-repellent properties. Their potential applications range from lab on a chip devices to self cleaning coating for clothes, glasses,... The actual strategy consist in mimicking superhydrophobic biosurfaces designing rough substrates out of hydrophobic materials. To achieve this goal both top-down and bottom-up approaches have been successfully developed: chemical synthesis of fractal surfaces [2], growth of carbon nanotube forests [3], deep silicon dry etching [4], see also [5] and references therein. We briefly recall the paradigm to account for superhydrophobicity. Two different wetting states can be observed on microstructured hydrophobic surfaces: (i) Wenzel state: the liquid follows the topography of the solid surface. Defining the surface roughness ζ as the ratio between the total surface area over the apparent surface area, the equilibrium contact angle of a liquid drop is given by cos(θ) = ζ cos(θ flat ),