iCosmos › The Hubble tension
The Hubble tension
Two ways of measuring today’s expansion rate H0 disagree. Compare the published values and see what each one implies for the age and size of the universe.
The interactive calculator loads here: change the redshift and the cosmological parameters (Ωm, ΩΛ, Ωr, w0, wa, H0) and the result and plot update instantly.
What it is
The Hubble constant H0 is the expansion rate of the universe today: a galaxy at a distance d recedes at about H0 d, as long as d is small compared with the size of the observable universe. It is quoted in kilometres per second per megaparsec (km s−1 Mpc−1). Because it sets the overall scale of time and distance, every age and every distance on this site is proportional to 1/H0 at fixed density parameters.
There are two broad routes to H0. The first measures it directly in the nearby universe with a distance ladder: geometric distances calibrate a type of star, those stars calibrate type Ia supernovae in the same galaxies, and the supernovae reach far enough into the smooth Hubble flow to measure the expansion rate. The second infers H0 from the early universe: the cosmic microwave background (CMB) fixes the physical densities and the angular size of the sound horizon very precisely, and a cosmological model, usually flat ΛCDM, then predicts what H0 must be today.
What the measurements say
The table below gives each value as the authors published it, with its uncertainty and the paper. The Planck 2018 CMB analysis gives 67.36 ± 0.54 km s−1 Mpc−1. The Atacama Cosmology Telescope DR6 combined with Planck, with CMB lensing and the DESI DR1 baryon acoustic oscillations added, gives 68.22 ± 0.36. Both assume flat ΛCDM. The SH0ES team, calibrating supernovae with Cepheid variable stars, measures 73.04 ± 1.04 and reports a 5σ difference from the Planck prediction. The Chicago–Carnegie Hubble Program, calibrating supernovae with the tip of the red giant branch observed with the Hubble and James Webb space telescopes, finds 70.39 ± 1.22 (stat) ± 1.33 (sys) ± 0.70 (σSN) and reports results consistent with ΛCDM. The first gravitational-wave standard siren, GW170817, gave 70.0 with an uncertainty of +12.0 and −8.0, independent of any distance ladder.
iCosmos does not compute its own significance for the disagreement: the measurements have different systematic uncertainties, and how to combine them is part of the scientific debate. The significance quoted above is the one the SH0ES authors state in their paper.
What changes when H₀ changes
Use the buttons in the table to set your model’s H0 to any published value; nothing else changes. The result cards above update at once. At fixed Ωm and ΩΛ, a higher H0 gives a younger universe and smaller distances, both in proportion to 1/H0. The worked example below shows the size of the effect for the classic model.
In a real analysis the other parameters move as well. The CMB constrains Ωmh2 and the angular scale of the sound horizon, so a CMB-based fit with a higher H0 would also have a lower Ωm. To explore that, change Ωm together with H0, or use the Planck and DESI presets in the inputs.
What supernovae can and cannot tell us
Type Ia supernovae are standardisable candles: after corrections for light-curve shape and colour, their peak brightness is nearly the same. On their own they measure relative distances, which fix the shape of the expansion history (how fast the expansion has accelerated) but not its scale. Their absolute magnitude and H0 are degenerate, so a supernova sample needs an external calibration, such as Cepheids or red giants in nearby host galaxies, before it can give H0.
The supernova plots on this page show this directly. Pantheon+ distances “as published” are calibrated with the SH0ES Cepheids, so with H0 near 67 the points sit about 0.2 mag below your model curve (0.18 mag for the Planck 2018 preset). That is close to 5 log10(73.0/67.4) ≈ 0.17 mag, the Hubble tension expressed in magnitudes; the rest comes from small differences in the shape of the curve. In “shape only” mode each dataset is moved by one constant, and changing H0 no longer moves the curve relative to the points.
Possible explanations
Proposed explanations fall into two groups: unrecognised systematic errors in one or more of the measurements (for example in the calibration of the distance ladder, or in how supernova host galaxies are treated), or physics beyond flat ΛCDM, usually acting before recombination so that the sound horizon is smaller than ΛCDM predicts. Independent methods, such as more gravitational-wave sirens and other distance-ladder calibrators measured with JWST, are being used to test both possibilities.
Formula
- Hubble law at small distance: v ≈ H0 d
- Hubble time: tH = 1/H0 = 977.8 Gyr / (H0 in km s−1 Mpc−1)
- At fixed Ωm, ΩΛ: t0 ∝ 1/H0 and DC(z) ∝ 1/H0
Worked example
For Ωm = 0.3 and ΩΛ = 0.7, H0 = 67.36 km s−1 Mpc−1 (Planck 2018) gives an age of 13.99 Gyr and a comoving distance to z = 1 of 3,433.3 Mpc. H0 = 73.04 km s−1 Mpc−1 (SH0ES) gives 12.91 Gyr and 3,166.3 Mpc.
The ratio of the two ages is 1.0843, exactly 73.04/67.36: an 8.4 % disagreement in H0 is an 8.4 % disagreement in every time and distance at fixed density parameters.
| Measurement | H0 (km s−1 Mpc−1) | Method | Source |
|---|---|---|---|
| Planck 2018 | 67.36 ± 0.54 | Cosmic microwave background (TT,TE,EE+lowE+lensing); assumes flat ΛCDM | Planck Collaboration 2020, A&A 641, A6, arXiv:1807.06209 |
| ACT DR6 + Planck + lensing + DESI DR1 BAO | 68.22 ± 0.36 | Cosmic microwave background (ACT DR6 + Planck, with ACT and Planck CMB lensing) combined with DESI DR1 baryon acoustic oscillations; assumes flat ΛCDM | Louis et al. 2025, JCAP 11 (2025) 062, arXiv:2503.14452 |
| SH0ES | 73.04 ± 1.04 | Distance ladder: Cepheids calibrating type Ia supernovae | Riess et al. 2022, ApJL 934, L7, arXiv:2112.04510 |
| CCHP (TRGB) | 70.39 ± 1.22 (stat) ± 1.33 (sys) ± 0.70 (σ_SN) | Distance ladder: tip of the red giant branch calibrating type Ia supernovae (HST + JWST) | Freedman et al. 2025, ApJ 985, 203, arXiv:2408.06153 |
| GW170817 standard siren | 70.0 +12.0 −8.0 | Gravitational waves from a neutron-star merger with an optical counterpart | Abbott et al. 2017, Nature 551, 85, arXiv:1710.05835 |
Common questions
What is the Hubble tension?
The disagreement between H0 measured directly in the nearby universe with Cepheid-calibrated supernovae (about 73 km s−1 Mpc−1) and H0 inferred from the cosmic microwave background assuming flat ΛCDM (about 67 to 68 km s−1 Mpc−1).
Which value of H₀ is correct?
This is not settled. The CMB values are precise but depend on the cosmological model; the distance-ladder values are direct but depend on how each rung is calibrated, and the two distance-ladder teams shown here (SH0ES with Cepheids, CCHP with red giants) currently report different central values.
Does the Hubble tension change the age of the universe?
Yes. At fixed density parameters the age is proportional to 1/H0, so the two families of values differ by about 8 % in age. Use the buttons above to see the age for each published value.
Why can’t supernovae alone measure H₀?
Their brightness gives relative distances. Without knowing their absolute magnitude, which has to be calibrated with nearby stars of known distance, the same supernova data fit any H0.
Why does iCosmos not show which side is “winning”?
Combining measurements with different systematic errors requires choices the authors of each analysis make carefully. iCosmos shows each published value with its own uncertainty and leaves the comparison to you.
Related calculators
- Age of the universe calculator
- Distance modulus calculator
- Luminosity distance calculator
- Hubble parameter H(z) calculator
- The full iCosmos cosmology calculator
About iCosmos
iCosmos was created by Mihran Vardanyan. Mihran Vardanyan is a cosmologist who earned his DPhil in cosmology at the University of Oxford, supervised by Joseph Silk and Roberto Trotta. He developed iCosmos during his research years at Oxford Astrophysics, where his work applied Bayesian model selection to the curvature and size of the Universe.
How to cite: Vardanyan, Mihran. iCosmos V4: Cosmology Calculator. https://icosmos.co.uk/. Include your access date, model parameters and exported data when reporting calculations.
iCosmos 4.5.0 · updated