Is the Hubble Tension a Local Hole in Space?

Imagine trying to measure the expansion of the universe from inside a giant cosmic bubble. That is the premise of a new arXiv Perspective by Indranil Banik, Harry Desmond, Vasileios Kalaitzidis and Sergij Mazurenko, posted on 3 February 2026. Their question is direct: could the Milky Way's address inside a vast underdensity of galaxies explain why nearby measurements of the Hubble constant disagree with the value inferred from the early universe? The paper presents no new telescope data; it is a synthesis of published evidence and model predictions.

A nine-percent disagreement that lives next door

The Hubble constant, H0, is the expansion rate of the universe today. From the cosmic microwave background (CMB), the ancient light released when the universe became transparent, a combination of Planck, the South Pole Telescope and the Atacama Cosmology Telescope gives H0 = 67.24 ± 0.35 km/s/Mpc, assuming the standard ΛCDM model of cold dark matter and dark energy. Local measurements, mostly using stars and supernovae within redshift z below 0.15, give a community consensus of 73.50 ± 0.81 km/s/Mpc. Redshift measures how much cosmic expansion has stretched light; z = 0.15 is the relatively nearby universe. The two numbers differ by about 9 percent, a tension above 7σ (sigma is a standard statistical measure; higher values mean stronger disagreement). Intermediate-redshift probes, from baryon acoustic oscillations to the ages of ancient stars, mostly agree with the CMB value. The disagreement seems to live only in our local neighborhood.

A hole about 300 megaparsecs in radius

The Perspective asks whether that neighborhood is special. It focuses on the Keenan-Barger-Cowie (KBC) void, a claimed underdensity of galaxies with a radius of about 300 megaparsecs; one megaparsec is about 3.26 million light-years. Galaxy counts in near-infrared and other wavelengths suggest we sit inside it. In redshift space, the void appears about 46 ± 6 percent underdense. Part of that apparent emptiness comes from redshift-space distortion: outflows from a void stretch redshifts and make the underdensity look deeper than it is. The actual underdensity may be about half as large. Even so, a simple continuity estimate used in the paper implies that the locally inferred recession scale, cz', should exceed the CMB-based H0 by 11 ± 2 percent.

That is the core mechanism. If we live in a void, galaxies around us drift outward faster than a smooth expansion would predict, and we also sit on a gravitational hill. Both effects add non-cosmological redshift. An observer who assumes a homogeneous universe would then infer too large a value of cz', even if the true background expansion rate is exactly the CMB value. The effect should fade with distance as the void's influence weakens.

Why the BAO ruler tilts at low redshift

The paper connects this idea to a second anomaly: the BAO distance scale. Baryon acoustic oscillations are frozen sound waves from the early universe, visible as a preferred clustering scale of about 147 comoving megaparsecs in galaxy maps (comoving distances remove the stretching of space). Astronomers use it as a standard ruler. When they compare the measured ruler to the prediction from the Planck cosmology, the ratio α_iso should be 1. Instead, low-redshift BAO data show α_iso below 1. The paper's void models, which were not tuned to BAO, predict α_iso ≈ 0.91 as z approaches 0 and a return toward 1 at higher redshift. The paper reports tensions of roughly 3 to 4σ depending on which BAO measurements and methods are combined, including about 2.9σ from DESI Data Release 2 alone.

Crucially, semi-analytic void models published earlier had already predicted this low-redshift BAO dip. Those models were constrained using the local expansion rate and galaxy counts, not BAO data. When compared with 20 years of BAO measurements, they reduce the tension substantially, to roughly 1.1–1.3σ in the paper's accounting for the larger compilation. They also predict that the inferred H0 should converge rapidly to the CMB value for z greater than about 0.2. Recent reconstructions of H0(z)—the H0 inferred from data in a narrow redshift slice—from BAO plus uncalibrated Type Ia supernovae (with their absolute brightness not fixed by a local distance ladder) show just such a downward trend, though the paper calls for a more quantitative comparison and the result remains model-dependent.

A tension that fades beyond z ≈ 0.2

The idea is not proven. The KBC void remains a claimed structure and is not yet firmly established. If taken at face value, its size and depth are about 6σ discrepant with ΛCDM simulations, which struggle to grow such a large void from the smooth initial conditions seen in the CMB. Explaining it may require faster growth of structure than in ΛCDM on scales larger than about 100 megaparsecs, perhaps through a modification of gravity. Bulk flows in the nearby universe, measured with the Cosmicflows-4 catalogue, appear larger than ΛCDM expects, but their significance depends on distance errors and selection effects. Some studies report anisotropy in the local cz'; a recent analysis finds none. And a local void is not the only way to produce a low-redshift BAO anomaly: a late-time change in dark energy could mimic some predictions.

What would prove a local void, or rule it out

Future observations could separate the scenarios. Fast radio bursts (FRBs), brief radio flashes, offer one test. Their pulses arrive later at lower frequencies because free electrons in intervening ionized plasma slow them down. That delay measures the electron column along the line of sight and is reported as the dispersion measure. If astronomers convert it to distance assuming a smooth universe while we actually live in a void, they will underestimate the distance and overestimate H0. One recent FRB-based estimate is around 76–77 km/s/Mpc, with uncertainties near 3 km/s/Mpc; the paper notes this is remarkably close to a rough void expectation. The kinematic Sunyaev-Zel'dovich effect, which uses CMB photons scattering off hot cluster gas to measure line-of-sight velocities, could probe the expected outflows. Redshift drift, the tiny change in a source's redshift over years, might one day detect the void's evolving pull, but distinguishing the void models from Planck cosmology would require a few millimeters-per-second precision over a decade.

A cosmic address, not a verdict

For now, the local void model is a coherent and testable story rather than a settled answer. It reframes the Hubble tension as a possible consequence of our place in the cosmos—a quiet, sparse patch whose gravity makes the nearby universe look faster than it really is—though the void itself may require changes to structure growth on very large scales. If that picture survives deeper galaxy maps, better velocity measurements and new radio surveys, our cosmic address may become part of the solution. If it does not, the tension will remain one of the sharpest clues that something in cosmology is still missing.