Inspired by Marshall-Olkin’s approach, multivariate distributions can be constructed through the use of exponential mixtures. In this paper, we propose an alternative hierarchical Archimedean copula, obtained from multivariate survival functions of multivariate mixed exponential distributions. The key element of our construction is that the latter are defined with a vector of mixing random variables, which follows a multivariate gamma distribution such as Kibble’s bivariate gamma distribution. After presenting the construction technique, properties of this new family of copulas are investigated, simulation algorithms are provides and illustrative examples are given. Risk aggregation and capital allocation under this newly proposed dependence structure are also examined.