Abstract: We discuss expectations of saturation physics for various observables in heavy ion collisions. We show how simple saturation‐inspired assumptions about particle production in heavy ion collisions lead to Kharzeev‐Levin‐Nardi model. Comparing this model to RHIC data on particle multiplicities we conclude that saturation effects may play an important role in particle production and dynamics at the early stages of Au − Au collisions already at RHIC energies. We then estimate the contribution of the initial state two‐particle azimuthal correlations to elliptic flow observable v2 in Au − Au collisions by constructing a lower bound on these non‐flow effects based on v2 obtained from the analysis of proton‐proton (pp) collisions.