Reanalysis of megamaser galaxies finds Hubble constant 69.7 ± 2.7 km/s/Mpc, shifting toward CMB value

A revised arXiv preprint reports that a fresh analysis of megamaser galaxies puts the Hubble constant at 69.7 ± 2.7 kilometers per second per megaparsec, noticeably lower than the earlier megamaser estimate and closer to the value inferred from the cosmic microwave background than to the leading distance-ladder result.

That matters because megamasers — galaxies whose central black holes are surrounded by bright, microwave-emitting disks — provide geometric distances and do not rely on the traditional Cepheid-and-supernova distance ladder. In the long-running debate over the “Hubble tension,” they have been one of the most important independent local cross-checks.

The paper, “A low value of H0 in tension with the distance ladder from Megamasers using peculiar velocity reconstruction,” is by Richard Watkins of Willamette University and Hume A. Feldman of the University of Kansas. It is a preprint posted to arXiv, originally submitted Aug. 6, 2026, and last revised Nov. 6, with the current PDF dated Nov. 9. The authors reanalyze the same six megamaser galaxies used in the Megamaser Cosmology Project study led by Geoffrey Pesce in 2020. Their main change is how they correct each galaxy’s redshift for peculiar velocity, the local motion of a galaxy on top of the overall cosmic expansion.

Instead of relying on older corrections, Watkins and Feldman use the McAlpine et al. 2025, or Manticore-Local, reconstruction of the nearby velocity field. That method combines constrained cosmological simulations with redshift-survey data to estimate how local gravitational flows nudge galaxies toward or away from us. For nearby objects, those motions can materially shift a redshift-based Hubble constant.

Using that approach, the authors write, “Our main result is that by using the recent velocity reconstruction of M25, we have obtained a Hubble constant of 69.7 ± 2.7 km/s/Mpc, closer to the CMB value although consistent with both the CMB and distance ladder values.”

The numerical shift is substantial. Pesce and colleagues reported 73.9 ± 3.0 in 2020, a result widely cited as an independent late-universe measurement in line with the higher local value. By comparison, the Planck satellite’s 2018 analysis of the cosmic microwave background gave 67.4 ± 0.5, while the 2026 H0DN distance-ladder result came in at 73.50 ± 0.81.

Just as important, the new preprint shows that the megamaser value changes significantly depending on how galaxy motions are handled. With no peculiar-velocity correction, the authors get 73.6 ± 3.0. Using Carrick et al. 2015 as in Pesce et al., they get 71.7 ± 2.7. Using Carrick “as intended,” they get 70.2 ± 2.6. With M25 and a fixed velocity dispersion of 150 kilometers per second, the result is 70.4 ± 2.8. Their M25 fiducial case lands at 69.7 ± 2.7.

That does not mean the Hubble tension is resolved, or even that the megamaser result by itself strongly conflicts with the distance ladder. The revised value is closer to Planck’s, but it still overlaps with both major camps within the stated uncertainties. Using the published errors, its difference from the H0DN result is about 1.35 standard deviations — a meaningful shift, not a decisive break.

The paper has already drawn technical criticism. Dan Scolnic, a cosmologist known for work on supernova cosmology, argued in public comments that the full M25 ensemble and covariance should be carried through the Hubble-constant calculation, rather than relying mainly on central velocity estimates plus per-galaxy dispersions. That criticism is methodological, not a settled refutation. But it underscores the main takeaway from this preprint: with a sample of only six galaxies, the megamaser estimate appears more sensitive to peculiar-velocity modeling than many readers may have assumed.