The Hubble Constant from Symbolic Regression of Cosmic Chronometers, BAO, and Three Independent Supernova Samples

We apply symbolic regression via PySR to cosmic chronometer (CC), BAO (SDSS + DESI DR1), and Pantheon+ supernova data to discover the cosmic expansion history H (z) with minimal theoretical priors. The data-driven search independently converges on a 4-parameter polynomial form H (z) = H0 + A z (z-B) (z²+C) with f (0) =0. Joint fitting yields H0 = 68. 0 +/- 0. 8 km/s/Mpc (68% CL), consistent with Planck 2018 at 1. 2 sigma and excluding SH0ES 2024 at 5. 0 sigma. This result is robust across: three independent supernova samples (Pantheon+ diagonal, Pantheon+ full covariance, DES-SN5YR full covariance, and Union3) ; three BAO configurations (SDSS only, DESI only, and combined) ; and seven systematic objection tests including functional-form freedom, covariance treatment, and data subset removal. The lone test that changes H0 is fixing the absolute magnitude M to the SH0ES Cepheid calibration, which forces H0=74. 4 with Delta chi² = +64 to +82 (8-9 sigma rejection). A direct LambdaCDM fit to the same data gives H0=67. 9, Omegaₘ=0. 32, joint chi²=1429. 4 --- nearly identical to the SR solution (chi²=1430. 6), confirming that the data-driven expansion history is competitive with the standard model. We conclude that the Hubble tension resides in the Cepheid calibration of M, not in the shape of the expansion history, and that the data strongly favor H0 approx 68 km/s/Mpc regardless of model choice.