The Hubble Tension

Two independent measurements of H₀ disagree at 5σ: Planck's CMB ladder gives 67, SH0ES's Cepheid ladder gives 73.

Suggested next → Dark Matter: Evidence and the Identification Gap · PHYS

The brief

For nearly a century, astronomers have argued over a single number: H₀, the rate at which the universe expands today. The old fights were about measurement error, and by the early 2000s they seemed settled. The new disagreement, sharpening since 2013, is stranger — two of the most careful methods we have simply do not agree. Read the expansion rate from the cosmic microwave background, the light left over from the infant universe, and you get about 67. Build it up instead from nearby stars and exploding supernovae, rung by rung out into space, and you get about 73. The gap is far too wide for either side's error bars to explain, and every improvement in precision has only made it harder to dismiss. This is the Hubble tension.

The two numbers come from opposite ends of cosmic history. The early-universe route runs through the standard model of cosmology, ΛCDM — ordinary matter, dark matter, and dark energy: the Planck satellite mapped the faint temperature ripples in the microwave background with exquisite precision, and, assuming that model is correct, those ripples fix the expansion rate at 67.4. The local route climbs a distance ladder: the geometric wobble of nearby stars calibrates Cepheids, pulsating stars whose rhythm reveals their true brightness; Cepheids in nearby galaxies calibrate Type Ia supernovae, which all explode at nearly the same brightness; and those supernovae, visible clear across the cosmos, give the expansion rate directly. The SH0ES team, led by Adam Riess, gets 73.0. The 8% gap between the two has survived every recheck.

There are only three ways out. Perhaps a subtle systematic error hides in one of the chains — but none has been found, and Riess's 2024 measurements with the James Webb telescope, confirming the Cepheid distances, make the local ladder look sound. Perhaps the answer is new physics: some ingredient missing from the early-universe model, such as a brief early burst of dark energy, that would nudge the CMB value upward — though every proposed fix tends to heal one dataset while straining another. Or perhaps it is a statistical fluke, a hope that fades as the gap stubbornly holds. The most intriguing recent hint comes from DESI, which in 2024 read the faint regular spacing imprinted on the distribution of galaxies and found a whisper that dark energy may itself be changing over time — which, if it holds, would point toward the new-physics door.

Why nowA genuinely independent method may yet settle it: standard sirens. When two neutron stars spiral together, their gravitational waves encode the distance directly, with no ladder to climb; pair that with the host galaxy's recession speed and H₀ falls out. The first such event, in 2017, landed loosely between the two camps, but once tens of them are logged the method should reach percent-level precision — and the giant detectors planned for the 2030s are being built partly for this purpose. Within a decade the verdict should be in: a missed systematic, real new physics, or a fluke that finally regresses to the mean. Rarely is a single number so load-bearing for cosmology.

Further readingCosmological Distances and the Hubble Tension (Riess, 2020 review). Modern Cosmology (Dodelson & Schmidt, 2nd ed., 2020). Planck 2018 Results VI: Cosmological Parameters (Planck Collaboration). The DESI 2024 BAO and Cosmology Results (DESI Collaboration).