The universe appears to be expanding faster than our current model predicts — and James Webb has now strengthened the case that the discrepancy is real, not just a flaw in Hubble’s measurements.

The Hubble tension is not a disagreement about whether the universe is expanding. It is a disagreement about how fast it is expanding, depending on which end of cosmic history is used to measure it.

Measurements anchored in the early universe, especially the cosmic microwave background as interpreted through the standard Lambda CDM cosmological model, point to a Hubble constant near 67 kilometres per second per megaparsec. Measurements built from the local distance ladder, using Cepheid variable stars and Type Ia supernovae, come out closer to 73.

That difference may sound small. In cosmology, it is not. It is large enough that it has become one of the central stress points in the current model. For years, one of the obvious questions was whether the local measurement might be wrong because the Hubble Space Telescope was seeing Cepheid stars in crowded regions where neighbouring stars blurred the light.

James Webb has now made that explanation harder to keep. With sharper infrared vision, Webb can separate stars in crowded fields better than Hubble could. In a 2024 paper in The Astrophysical Journal Letters, Adam G. Riess and colleagues reported that JWST observations reject unrecognised crowding of Cepheid photometry as an explanation for the Hubble tension at 8 sigma confidence.

What Webb actually tested

The local distance ladder works by building a chain. Cepheid variable stars are useful because their pulsing periods reveal their intrinsic brightness. If astronomers know how bright a Cepheid really is, and how bright it appears from Earth, they can estimate its distance. Cepheids then calibrate Type Ia supernovae in nearby galaxies, and those supernovae allow distances to be measured much farther out.

Hubble was central to that chain. Its measurements underpinned the SH0ES team’s 2022 result, also published in The Astrophysical Journal Letters, which reported a local Hubble constant of about 73 kilometres per second per megaparsec, with an uncertainty near 1.

The worry was technical but serious. If Hubble’s images of distant Cepheids were contaminated by nearby unresolved stars, those Cepheids could appear brighter than they really were. Brighter would mean closer. If enough Cepheids were affected in the right direction, the local measurement of the expansion rate could be pushed upward.

Webb was well suited to test that. Its infrared instruments can observe Cepheids with finer resolution and less interference from dust. The point was not that Webb would make the tension vanish by taking a prettier image. The point was that it could examine one of the most plausible measurement biases in the Hubble data.

The escape route narrows

Riess and colleagues had already published a 2023 JWST check titled “Crowded No More”, testing the accuracy of the Hubble constant with high-resolution Webb observations of Cepheids. The 2024 paper extended that argument and sharpened the conclusion: the kind of unrecognised crowding needed to remove the Hubble tension was not showing up in the Webb comparison.

That does not mean every possible systematic error is dead. Cosmology is a long chain of inference, and each link has to be examined. Cepheid calibration, supernova standardisation, host-galaxy effects, dust, sample selection and alternative distance indicators all remain part of the debate.

But one simple version of the sceptical argument has become less persuasive. If Hubble had been fooled by crowded Cepheid fields, Webb should have seen the mismatch. Instead, the Webb observations supported the Hubble-based distance ladder closely enough to strengthen the case that the discrepancy is not just an artefact of Hubble’s resolution.

The other side of the tension

The early-universe value comes from a very different route. The European Space Agency’s Planck mission measured the cosmic microwave background, the relic radiation from the young universe. When those data are interpreted using the standard Lambda CDM model, the inferred present-day expansion rate is lower than the local distance-ladder value. The Planck 2018 results, published in Astronomy & Astrophysics, remain a key reference point for that lower value.

That is why the tension matters. It is not one telescope disagreeing with another in the same measurement system. It is the young universe, interpreted through a successful model, disagreeing with the nearby universe measured through stars and supernovae.

The standard model has not failed wholesale. Lambda CDM still explains a remarkable amount: the cosmic microwave background, large-scale structure, the broad history of expansion and the abundance of light elements. The Hubble tension is troublesome precisely because it sits inside a model that otherwise works so well.

What not to conclude

The Webb result is not evidence that a new component of physics has been identified. It does not tell us that dark energy behaves differently, that early dark energy existed, that gravity changes, or that any particular alternative model is correct. It does something narrower and still important: it tests one major observational concern and finds that it is unlikely to explain the discrepancy.

That is a different kind of scientific progress. It does not produce a dramatic answer. It removes an easier answer.

The case for a real tension is stronger when independent checks point the same way. Webb’s Cepheid work does not settle the issue alone, but it reinforces the local ladder rather than undermining it. If the 73-ish value is not a Hubble crowding artefact, then the gap with the 67-ish early-universe value becomes harder to dismiss as a technical nuisance.

A disagreement that survived a better telescope

There is a quiet irony here. The James Webb Space Telescope was not built to protect the Hubble Space Telescope’s legacy. Yet in this case, Webb has given Hubble’s distance measurements one of their strongest external checks. The newer telescope did not make the old discrepancy disappear. It made the simplest measurement-error explanation look weaker.

That is why the result has weight. A tension that vanishes with better data is usually a measurement problem. A tension that survives better data becomes more interesting.

For now, the universe appears to be expanding faster nearby than the standard early-universe model predicts for today. Webb has not told cosmologists why. But it has made it harder to say the whole problem was just Hubble seeing crowded stars too poorly.