The Axidental Universe

Random axion theories with several hundred fields have enormous numbers of distinct meta-stable minima. A small fraction of these have vacuum energy compatible with current measurements of dark energy. The potential also contains regions suitable for inflation, and gives rise to a natural type of dark matter. First-order phase transitions from one minimum to the vicinity of another play the role of big bangs and produce many bubbles containing evolving Friedmann-Lemaître-Robertson-Walker universes. The great majority either collapse in a tiny fraction of a second, or expand exponentially forever as empty, structureless universes. However, restricting to those bubble universes that form non-linear structure at some time in their history we find cosmologies that look remarkably similar to ours. They undergo about 60 efolds of inflation, making them flat, homogeneous and isotropic, and endowing them with a nearly scale-invariant spectrum of primordial density perturbations with roughly the observed magnitude and tilt. They reheat after inflation to a period of radiation domination, followed by matter domination with roughly the observed abundance, followed by vacuum energy domination at roughly the observed density. These features are largely insensitive to the dimensionful and dimensionless parameters of the theory, which can be set to the grand unified scale and order one respectively. In our benchmark model we assume the number of high-scale contributions to the axion potential is not much larger than the amount of axions, and that there is a single field direction which is left massless by these contributions. The small value of dark energy ultimately comes from non-perturbative gravitational effects, giving ρDE ≈ Λ4 e-𝒪(1) × M_Pl/f, where f ≈ Λ ≈ 10-2MPl. Therefore, random axion landscapes can account for many of the apparently tuned features of our universe, including its current enormous size, age, and tiny energy densities compared to the scales of fundamental physics.