If you ask the early universe what the Hubble constant is, it gives you one answer. If you ask the local universe, it gives you a different one. Both measurements have been made with independent techniques, refined with progressively larger datasets and more sophisticated systematic error analyses, and the disagreement between them has grown rather than shrunk as precision has improved. This is the Hubble tension, and it is either the largest systematic error in the history of precision cosmology or evidence that the standard model of the universe is missing something fundamental.
The two numbers at the center of the tension are 67.4 and 73.0 kilometers per second per megaparsec. The first comes from the Planck satellite's measurement of the cosmic microwave background. The CMB encodes the conditions of the universe at 380,000 years after the Big Bang, and fitting the standard cosmological model (ΛCDM) to the CMB's temperature and polarization anisotropies yields a precise value of the Hubble constant consistent with 67.4. The second comes from what is called the distance ladder: measuring the distances to nearby Cepheid variable stars using parallax, using those Cepheids to calibrate the distance to galaxies hosting Type Ia supernovae, then using the supernovae as "standard candles" to measure distances to galaxies far enough away that their recession velocities directly measure the Hubble constant. The Hubble Space Telescope's SH0ES program, led by Adam Riess (Nobel laureate 2011), has refined this method for two decades. Its result: 73.0, with an uncertainty of about 1 percent.
Why it might be systematic error
The obvious first hypothesis is that one or both measurements contain a systematic error — an unaccounted bias in the calibration. For the CMB, the error would have to be in the assumed cosmological model: if ΛCDM is wrong in some subtle way, the derived Hubble constant would shift. For the distance ladder, it could be in the Cepheid calibration (dust reddening, metallicity effects, sample selection), in the supernova calibration (host galaxy environment effects, intrinsic luminosity spread), or in both. Webb observations have re-observed many of the Cepheids and supernovae in SH0ES's sample, with higher resolution and sensitivity than Hubble, partly to check for contamination of the Cepheid photometry by neighboring stars — a plausible systematic in Hubble's resolution. The Webb analysis, published in 2023 by the SH0ES team, confirmed the Cepheid distances and the 73 km/s/Mpc result. The tension did not shrink.
What new physics could cause it
If the tension is real and not systematic, the standard cosmological model needs to be extended. The CMB measurement assumes the universe's composition and behavior for 13.8 billion years — if dark energy is not a fixed cosmological constant but evolves over time (as some recent DESI survey results hint), the inferred Hubble constant from the CMB shifts. Another possibility is "early dark energy" — a brief burst of additional energy density in the early universe that would change the sound horizon scale in the CMB, shifting the derived H0 upward. Neither explanation fits comfortably with all constraints. The Hubble tension is the most precisely defined anomaly in cosmology — it sits at 5 sigma, far enough above the noise that dismissing it as a fluctuation requires extraordinary justification. Its resolution, when it comes, will say something definitive about what the universe is made of and what rules it follows. The history of cosmology contains several such moments: the discovery of the CMB resolved the steady-state vs. Big Bang debate; the supernova measurements of 1998 revealed dark energy. Both were surprises. The Hubble tension may be the next one — a discrepancy between two numbers that turns out to be evidence not for an error but for a piece of the universe that no current model has accounted for. Cosmologists are aware of the precedent and are proceeding accordingly.
