At the present time, cosmology is presenting us with a wealth of data which can only be explained by invoking new physics operating in the very early universe. I will introduce three paradigms for the evolution of the early universe: inflation, a cosmological bounce, and an emergent scenario. I will argue that in light of arguments from quantum gravity and string theory the inflationary paradigm is severely constrained.
I will then discuss attempts to obtain an emergent cosmology making use of fundamental principles of superstring theory, in particular matrix theory.