We provide extension procedures for nonlinear expectations to the space of all bounded measurable functions. Based on a nonlinear version of the Daniell-Stone theorem we deduce a robust Kolmogorov extension theorem which is then used to extend nonlinear kernels to an infinite dimensional path space. As an llustration, we construct nonlinear Markov chains in discretetime. In the second part of the talk we construct sublinear semigroups and discuss the link to non-linear expectations and PDEs. The talk is based on joint work with Robert Denk and Max Nendel.