Ultra-light axions (m_a<10^{-18} eV), motivated by string theory, can be a powerful probe of the energy scale of inflation if they exist as a sub-dominant component of the Dark Matter. In contrast to heavier axions the isocurvature modes in the ultra-light axions can coexist with observable gravitational waves. Here it is shown that existing (2005) large scale structure constraints severely limit the parameter space for axion mass, density fraction and isocurvature amplitude. It is also shown that radically different CMB observables for the ultra-light axion isocurvature mode additionally reduce this space. The results of a new, accurate and efficient method to calculate this isocurvature power spectrum are presented, and can be used to constrain ultra-light axions and inflation.

I will also present preliminary results of constraints to this model using up-to-date cosmological observations, which verify the above picture. The parameter space is interesting to explore due to a strongly mass dependent covariance matrix, motivating comparisons between Metropolis-Hastings and nested sampling. Finally I discuss fine-tuning and naturalness in these models.