JWST was supposed to referee the Hubble tension and instead exposed how much the answer depends on analysis choices. The SH0ES team reports that JWST observations of Cepheids in supernova host galaxies confirm the HST photometry with no crowding bias, sustaining H0 = 73 (Riess et al.; IOPscience 2024-2025). The Chicago-Carnegie Hubble Program, observing with the same telescope and largely the same anchor (NGC 4258), reports JWST-only values of H0 = 68.81 +/- 1.79 (stat) from TRGB and 67.80 +/- 2.17 from the JAGB method, in excellent agreement with Planck, while its combined HST+JWST TRGB analysis gives 70.39 (arXiv:2408.06153, arXiv:2408.03474). The same instrument, in the same year, supports both sides of the tension.
The split traces not to the rungs but to the priors and the sample: CCHP and SH0ES distance measurements to the same galaxies agree at the 1 percent level across Cepheids, TRGB, and JAGB, yet the derived H0 differs by 4 to 5 percent. Riess et al. argue the CCHP low values are explained by the selection of the supernova calibrator subsample, which uses fewer SN hosts and so inherits different statistics; Freedman et al. argue the SH0ES result inherits Cepheid-specific systematics and that the method spread itself is the message. Each team's H0 thus carries an irreducible contamination from its analysis priors: which calibrators to admit, which anchor to trust, which method variance to propagate. ΛCDM offers no guidance, because it requires one true H0 and treats the spread as someone's mistake.
The standing is an uncomfortable stalemate at the heart of precision cosmology: the distance measurements agree, the cosmological conclusions do not, and the disagreement lives in choices the data alone cannot arbitrate. More JWST calibrators and independent geometric anchors will shrink the freedom, but the episode already shows the local H0 is not yet a single settled number.