We define a notion of Ricci curvature in metric spaces equipped with a measure or a random walk. For this we use a local contraction coefficient of the random walk acting on the space of probability measures equipped with a transportation distance. This notions allows to generalize several classical theorems associated with positive Ricci curvature, such as a spectral gap bound (Lichnerowicz theorem), Gaussian concentration of measure (Lévy–Gromov theorem), logarithmic Sobolev inequalities (a result of Bakry–Émery theory) or the Bonnet–Myers theorem. The definition is compatible with Bakry–Émery theory, and is robust and very easy to implement in concrete examples such as graphs. To cite this article: Y. Ollivier, C. R. Acad. Sci. Paris, Ser. I 345 (2007).