Two-Loop EFT Galaxy Power Spectrum Matches PT Challenge Simulations to 0.1 Percent up to k=0.85 h/Mpc

The calculation incorporates two-loop renormalization conditions, higher-derivative operators, stochastic terms, and time-sliced perturbation theory IR resummation, all interfaced with the CLASS-PT renormalization scheme. This extends prior one-loop analyses by systematically removing redundant operators and deriving the full set of counterterms needed for consistent ultraviolet subtraction at this order. The resulting model is directly portable to existing pipelines used for BOSS and eBOSS analyses.

Comparison against the PT Challenge suite demonstrates unbiased recovery of the linear power spectrum amplitude with three times smaller error bars on sigma8 than linear theory and a 40 percent improvement over one-loop results, even under conservative priors. The shape information retained between 0.4 and 0.8 h/Mpc directly tightens constraints on any new physics that alters the matter power spectrum in that window.

Weak-lensing surveys such as Euclid, LSST, and Roman will cross-correlate galaxy positions with CMB lensing or shear maps in 2x2pt analyses. The two-loop model reduces theoretical systematics that currently limit the k-range usable for these measurements, thereby increasing the effective volume probed and sharpening tests of ultra-light axion dark matter sub-components around 10^-24 eV. Forecasts that previously relied on one-loop modeling will require revision to capture the full statistical power of the upcoming data releases.

Next steps include public release of the two-loop kernels, joint analyses with bispectrum measurements, and full-shape fits to DESI Year-1 and Euclid early data that marginalize over the expanded 21-parameter space while retaining the reported precision gains on sigma8.