Cellular metabolism consists of a complex network of chemical reactions, which provide the cell with a reliable supply of energy and building blocks. Optimization approaches based on steady state approximations have been increasingly successful at predicting how individual microbial species allocate available resources under different environments and perturbations. Possible extensions of current approaches that incorporate interactions between different species can help understand natural microbial ecosystems and engineer artificial ones, with applications ranging from metabolic engineering to the human microbiome.