TDCOSMO 2026: A 3% measurement of the Hubble constant from strong lensing and stellar kinematics

TDCOSMO Collaboration, Simon Birrer, Michele Cappellari, Frédéric Courbin, Christopher D. Fassnacht, Pierre Fleury, Joshua A. Frieman, Aymeric Galan

Oct 10, 2026arXiv:2610.14781v1

astro-ph.CO

Scorecard· 16/16

8.0/10 impact

Most precise conservative (MSD-marginalized) lensing H0 to date at 3%, blind and heavily cross-checked, directly relevant to the Hubble tension, though an evolutionary update of TDCOSMO-2025.

Abstract

We present cosmological constraints from 11 strongly lensed quasars with measured time delays (hereafter, the TDCOSMO-2026 sample). In addition to expanding the sample, we improve many aspects of our blind analysis with respect to our 2025 analysis. Among other improvements, spatially resolved kinematics are available for nine systems, including six with JWST-NIRSpec, compared to one in 2025; for three lens models, JWST-NIRCam images now replace HST; axisymmetric dynamical modeling is applied directly to the spatially resolved data, instead of an axisymmetric correction factor applied to spherical models; and the line-of-sight convergence is now based on the Euclid Flagship simulations instead of the Millennium Simulation. Our baseline and primary result is based on the TDCOSMO-2026 sample, in combination with a prior on $Ω_{\rm m}$ from the DES-Dovekie supernova sample. In the flat $Λ$ cold dark matter ( $Λ$ CDM) cosmology, we find $H_0=71.1^{+2.4}_{-2.3}$ km.s $^{- 1}$ .Mpc $^{- 1}$ , a 3.3% uncertainty that accounts for the mass-sheet degeneracy. The SLACS and SL2S samples are in excellent agreement with the TDCOSMO-2026 sample, improving the precision on $H_{0}$ in flat $Λ$ CDM to 3.0%. Using the DESI-DR2 baryon acoustic oscillations (BAO) likelihood instead of DES-Dovekie yields fully consistent results. We also present constraints in the open $Λ$ CDM, $w$ CDM, $w_{0} w_{a}$ CDM, and $w_{φ}$ CDM cosmologies; the TDCOSMO $H_{0}$ inference is robust and consistent across all of them. The combination of the TDCOSMO+SLACS+SL2S likelihood with those from DESI-DR2 and DES-Dovekie disfavors $w = - 1$ at the $2.3$ - $2.5 σ$ level in all three dark energy extensions that we tested. We find, e.g., $w_0=-0.88^{+0.04}_{-0.05}$ in $w_{φ}$ CDM and $w_0 = -0.85^{+0.07}_{-0.06}$ in $w_{0} w_{a}$ CDM, while $w_a = -0.48^{+0.46}_{-0.44}$ remains consistent with zero.

AI Impact Assessments

(1 model)

Core Contribution

This is the TDCOSMO collaboration's 2026 milestone paper. It gives a blind hierarchical cosmographic analysis of 11 lensed quasars with time delays, and it reports H0 = 71.1 (+2.4/−2.3) km/s/Mpc in flat ΛCDM, a 3.3% uncertainty. Adding SLACS and SL2S non-time-delay lenses gives 3.0%.

The headline is the precision. The result keeps the mass-sheet degeneracy (MSD) as a free population-level parameter (λ_int) constrained by kinematics. Earlier "assertive" analyses reached 2% only by assuming power-law or NFW+stars profiles. The conservative approach was 8% in TDC4 and 5% in TDCOSMO-2025. So this paper is roughly the first conservative, MSD-marginalized lensing result that is competitive with the assertive ones.

The main methodological advances are:

spatially resolved JWST-NIRSpec/KCWI/MUSE kinematics for 9 of 11 lenses (about 229 bins), versus one lens in 2025;

direct 2D axisymmetric Jeans modeling (JAM) with joint sampling of inclination, anisotropy and intrinsic shape, replacing a statistical correction applied to spherical models;

conditioning D_Δt on the power-law slope γ_pl, which preserves the D_Δt–γ_pl correlation and avoids double-counting the slope uncertainty;

JWST-NIRCam lens models for three systems;

Euclid Flagship-based line-of-sight convergence, with the exact κ_d, κ_s and κ_ds terms rather than κ_ext ≈ κ_s.

The paper also reports constraints in open ΛCDM, wCDM, w0wa and w_φ CDM. Combined with DESI-DR2 and DES-Dovekie, these disfavor w = −1 at 2.3–2.5σ. The time-delay distance acts as an absolute anchor for BAO, which gives r_d = 140.7 ± 4.3 Mpc without CMB input.

Methodological Rigor

Rigor is high by the standards of the field.

The analysis is blind (median removal), with a formal unblinding, five internal reviewers and two code reviewers. Baseline choices were fixed before unblinding.

The paper reports goodness-of-fit checks for D_Δt (χ² = 10 for 11 lenses) and for the kinematics.

It runs one-at-a-time analysis variations: Ωm prior, external samples, scatter priors, q_obs definition, informative MaNGA priors, cylindrical vs spherical alignment, and three masking schemes. All shifts are below 1σ, with a maximum of +0.7σ.

It discusses selection effects and the SLACS/SL2S sample-matching cuts, and it keeps the time-delay-only result as primary because the external lenses carry an assumption of a shared population.

It is candid about caveats. The λ_int scatter is now detected at about 0.09 and is larger (about 0.13) when external lenses are included. This hints at a population difference or residual modeling systematics. The anisotropy is largely prior-driven. The result has a mild upward sensitivity of +0.5σ to the anisotropy prior.

Limitations in rigor:

Systematics are declared "subdominant" without a quantitative per-source error budget. The paper itself says this is future work.

The result depends on the assumption that λ_int is drawn from a Gaussian population shared across lenses, and on a linear dependence on R_eff/θ_E.

Several inputs are only summarized here and detailed in companion papers (lens modeling, κ_ext, kinematics), so the paper cannot be fully evaluated on its own.

The variations share the same data, so they test stability but not independence.

The wCDM and w0wa results stem mostly from DESI and DES-Dovekie, with lenses adding little. The FoM rises only 17%.

Potential Impact

H0 is a central problem in cosmology, and the lensing result sits between Planck (1.6σ below) and the local distance ladder (within 1σ). At 3% it is among the most precise late-universe probes and independent of the distance ladder, SN calibration and CMB. This gives it high relevance in the Hubble-tension debate, for which a companion PRL is also prepared. The result does not adjudicate the tension, since it is consistent with both sides, but it narrows the room for ladder-specific systematics. The paper also shows that a conservative MSD treatment need not give up precision, which will guide the analysis of future samples from Euclid, Rubin and Roman. The methods (conditioned D_Δt, 2D JAM, Euclid-based κ_ext, hierarchical inference with external lenses) form a reusable template. The 2.3–2.5σ dark-energy hint is mostly inherited from DESI and DES-Dovekie and adds little independent information.

Timeliness

The paper is very timely. It arrives amid Hubble-tension debates, DESI DR2 evolving dark energy hints, and new JWST kinematic data. It also anticipates large lens samples from LSST and Euclid.

Strengths and Limitations

Strengths:

It is the strongest and most mature time-delay cosmography result to date.

It is blind and conservative, with unusually extensive robustness testing.

It makes extensive use of new JWST data.

The central value is consistent with TDCOSMO-2025 and TDCOSMO-1, which supports stability.

Limitations:

It is an incremental (though substantial) update of the 2025 analysis.

11 lenses limit the precision, and the λ_int scatter sets a floor on how far it can improve.

It does not by itself resolve the tension.

Population-homogeneity and parameterization assumptions are not fully tested.

Its heavy reliance on companion papers and its long collaboration-style presentation hurt standalone clarity somewhat.

Reproducibility is moderate. Likelihoods are promised on acceptance, the pipeline uses public tools (lenstronomy, hierArc, jampy), but the data are drawn from many companion papers.

Overall

This is a high-impact, field-defining measurement in time-delay cosmography. It is moderate in novelty because it is an evolutionary step, and very strong in rigor, with high visibility in the Hubble tension debate.

Rating:8/ 10

Significance 8Rigor 8.5Novelty 6Clarity 6.5

Generated Oct 14, 2026

Comparison History (0)

No comparisons yet.

Scorecard· 16/16

8.0/10 impact

Most precise conservative (MSD-marginalized) lensing H0 to date at 3%, blind and heavily cross-checked, directly relevant to the Hubble tension, though an evolutionary update of TDCOSMO-2025.