Famous study claiming the Universe is anisotropic found to be in error

Full text

Here in our Universe, we’ve drawn the conclusion that it’s been expanding and cooling for 13.8 billion years: ever since the hot Big Bang first began. In all directions, the same cosmic structures emerge: stars, galaxies, groups and clusters of galaxies, a network of interconnected filaments, with vast cosmic voids separating these matter-rich structures. At distances near and far, and in all directions and all locations, the Universe appears not identical, but similar: with the same densities, galaxy counts, and types of structures found everywhere.

Our cosmological picture, however, only makes sense — and exhibits self-consistency — if the Universe is both homogeneous and isotropic: the same in all locations and the same in all directions. The underlying equations we use to govern the expanding Universe on the largest of cosmic scales, the Friedmann equations, require both of these assumptions to be true.

Thus far, the large-scale structure data seems to agree with these assumptions, including from the largest surveys of all: the 2dF galaxy redshift survey, the Sloan Digital Sky Survey, and the Dark Energy Spectroscopic Instrument (DESI) survey have all supported this consensus picture. However, in a new study published in Nature at the end of June 2026 , coauthors Francesco Sylos Labini and Marco Galoppo argue that the DESI data actually supports an anisotropic Universe, making for some rather splashy headlines .

The big problem is that the study is flawed, and the data doesn’t indicate anisotropy at all, as cosmologist Till Sawala has shown in a follow-up paper . Here’s what’s at stake, and what the Universe is actually telling us.

In the aftermath of inflation, signatures are imprinted onto the Universe that are unmistakably inflationary in origin. While the CMB provides an early-time “snapshot” of these features, that’s just one moment in history. By probing the large variety of times/distances accessible to us throughout cosmic time, such as with large-scale structure, we can obtain information that would otherwise be obscure from any single snapshot: an array of snapshots from all different cosmic epochs.

Credit : Caltech/Robert Hurt(IPAC)

When we look out at the Universe, what we’re seeing isn’t necessarily what your mind might intuit. When we look at a:

star,

galaxy,

group of galaxies,

or a cluster of galaxies,

we’re getting a snapshot of that object: how it was at one particular moment in time. Specifically, the moment of that snapshot shows us the object as it was when the now-arriving light first left that object, modified only by the effects of the environment(s) it passed through on its journey to our eyes and the expansion of the Universe that stretched that light’s wavelength.

However, because our “snapshots” contain objects at a wide variety of distances — including stars within our own galaxy, galaxies in the nearby Universe, and ultra-distant galaxies and groups/clusters of galaxies often all appearing in the same field-of-view — we’re actually getting disparate snapshots of objects at a variety of epochs throughout cosmic history. More distant objects appear as they were when the Universe was younger, hotter, denser, and less evolved, while more nearby objects appear as they are closer to today: when the Universe is older, colder, sparser, and more evolved.

And in order to make sense of what we’re seeing, we have to take all of those age and evolutionary effects into proper account.

This JWST view shows the five main galaxies of Stephan’s Quintet, and was one of the first 5 science images released by the JWST team on July 12, 2022. The galaxy on the left is only about ~15% as distant as the other four members of the quintet, while the spike-rich stars are within our own Milky Way. Far beyond, in the background, are galaxies and clusters of galaxies at all different distances: scores of times farther away than the quintet.

Credit : NASA, ESA, CSA, and STScI

If we do account for all of those effects properly, however, including the effects of:

our cosmic motion relative to the Hubble flow,

the expanding volume (and decreasing density) of the Universe over time,

the evolution of galaxy masses and clustering,

the limitations of what our observing telescopes are sensitive to in terms of brightness and wavelength,

and how to properly match up what simulations predict with what observations indicate,

then we should see that, on the largest of cosmic scales, the Universe really is the same everywhere and in all directions. That includes:

equal number densities of galaxies in all locations on the sky,

equal fractions of galaxies seen rotating in all directions,

equal populations of galaxies of different shapes/morphologies,

and a lack of obvious large-scale structures on larger scales than the (comoving) equivalent of 1.4 billion light-years by the present day.

If we zoom out sufficiently, on the largest of cosmic scales, even the massive and impressive amounts of structure that we see all across our Universe appear to blend together, giving way to a seemingly uniform background of galaxies clustered together in similar cosmic patterns.

This simulation shows the large-scale structure of the Universe on the scales of the actual cosmic horizon: approximately 93 billion light-years across. At this level, even the largest cosmic structures are reduced to mere blips, showcasing just how uniform, smooth, and equal the Universe is on the largest of cosmic scales.

Credit : The Millennium Simulation, V. Springel et al.; Modifications: E. Siegel

This underlying assumption — of the Universe being homogeneous and isotropic — is vital to even applying the very equations that led us to predict the expanding Universe. If it’s not true, then the foundation that we’ve built our cosmological picture atop is rotten to the core, and everything that we’ve thought up until this point needs to be re-examined. If the Universe turns out not to be homogeneous and isotropic, then well-established facts like:

the expanding Universe,

the hot Big Bang,

the existence of dark matter and dark energy,

and cosmic cornerstones like the CMB, the growth of large-scale structure, and the abundance of the light elements,

must all be thrown into doubt as well.

Of course, we don’t just believe in our underlying assumptions and hope that they hold true; we test them as rigorously as possible. We run large numbers of simulations where we design “mock universes,” or cosmic simulations that start off with the known laws of gravity, our suspected ingredients, and a set of initial conditions, and see what types of structures quantitatively emerge. We then compare the results from those simulations with the Universe that we actually observe through telescopes and surveys, and see if they’re consistent with one another. If they are, the assumptions seem consistent with our cosmic pictures, but if not, perhaps the core of cosmology , the cosmological principle itself , needs to be re-examined.

The inferred difference in motions from a variety of properties of galaxy clusters in different directions across the sky, including X-ray, brightest cluster galaxy, and Sunyaev-Zel’dovich effects. This effect was suspected by some to point towards a cosmic anisotropy, but subsequent data from eROSITA ruled out that possibility at overwhelming significance with superior observations.

Credit : K. Migkas et al., A&A, 2021

That’s precisely what we do whenever we take superior large-scale structure data from a new survey. You’ll want to make sure that:

there are no hemispherical asymmetries in galaxy counts, morphologies, or rotational directions,

the predicted galaxy distribution is consistent with the observed distribution to the limits of your survey’s observational capabilities,

that the local galaxy distribution, the intermediate-distance galaxy distribution, and the very distant observed distributions are all consistent with our cosmological model,

and that there’s no unexpected evolution, no indicators that dark matter or dark energy are absent, and no significant departures from what simulations predict.

If any of these observed effects run counter to what your best simulations predict, then you’ve got an anomaly on your hands: one that merits further attention.

Many such anomalies were claimed in the past, but were discovered to be artifacts of small sample sizes (that went away when larger samples were obtained), of incorrect analyses of the data (such as the evidence for quantized redshifts of quasars, which was thoroughly refuted with the advent of SDSS data), or of incorrect calibrations (where the correct calibration caused the purported anomaly to disappear). We don’t just leap whenever a claim for violation of cosmic isotropy gets made; we verify it to make sure it’s robust, rather than reflective of an error in analysis.

This slice of the DESI data maps celestial objects from Earth (center) to billions of light years away. Among the objects are nearby bright galaxies (yellow), luminous red galaxies (orange), emission-line galaxies (blue), and quasars (green). The large-scale structure of the universe is visible in the inset image, which shows the densest survey region and represents less than 0.1% of the DESI survey’s total volume.

Credit : Claire Lamman/DESI collaboration

Today, it’s routine for observers to conduct these cosmic tests and publish the results whenever a new survey result is obtained. That’s why it’s so interesting — yet not necessarily compelling — that a team claimed to discover exactly such an anisotropy using the latest, most comprehensive large-scale data that we have: Dark Energy Spectroscopic Instrument (DESI) data . Now, the DESI team themselves, in their original data releases, did this analysis and found that everything was consistent with our standard cosmological assumptions, and only when they folded in extra pieces of information (about the CMB and supernova data), did small tensions arise between things like neutrino masses and the question of whether dark energy was a constant or not. The big assumptions, like anisotropy and homogeneity, were only validated, according to their analyses.

However, after the data releases came out, that DESI data then became public, where anyone in the world is free to use and analyze it. Two scientists, Francesco Sylos Labini and Marco Galoppo, then took that DESI data and analyzed it themselves, looking to test the notion of cosmic isotropy. “Is the Universe really the same in all directions,” they asked, especially on scales greater than 1.4 billion light-years, where it should evidently be the same?

What they found was surprising. Using a statistical check known as the angular distribution of pairwise distances, they looked for directional correlations in how likely you were to find a galaxy located a specific distance away from any other galaxy. They found anisotropy signals that were far greater than were expected: at greater than 99.7% significance and on scales of more than 3 billion light-years.

This animation of DESI’s 3D map of the large-scale structure in the Universe, the largest such map to date, was created with the intention of studying dark energy and its possible evolution. However, although they found evidence for dark energy evolving, that’s likely due to the assumption that it’s dark energy’s evolution that’s causing the discrepancies in the data compared to our standard cosmological model. An independent analysis of DESI data led to a cosmic questioning of whether the Universe exhibits isotropy or not: an even more dubious (but revolutionary-if-true) claim.

Credit : DESI Collaboration/DOE/KPNO/NOIRLab/NSF/AURA/R. Proctor

If true, that claim would be revolutionary, indicating that one of our prime, foundational assumption about the Universe — that is really is the same in all directions — may actually conflict with the Universe as we observe it. However, there’s a common way, whenever new large-scale structure observations come out, that scientists have regularly fooled themselves into believing there was an anisotropy because they failed to properly calibrate what they were doing: ignoring the peculiar motion of the galaxies being observed throughout the Universe. (In astrophysics circles, this is sometimes known as the Kaiser rocket effect .)

Of course, in the literature, such claims get made all the time, as scientists frequently seek to challenge foundational assumptions. Examples of claims include:

local underdensities,

source count dipole anomalies (too many objects in one direction coupled with too few objects in the opposite direction),

gigaparsec-scale patterns ,

and are almost always examples of motivated reasoning. These claimed structures are usually shown, upon a further, independent analysis, to be consistent with random fluctuations, to be statistically insignificant, or to be perfectly compatible with predictions of the standard cosmological picture. This latest claim is a very strong one, contending that the clustering of bright sources in the DESI data shows large breakdowns and departures from traditional isotropy.

This graph shows a comparison of galaxy distributions in the DESI DR1 data (left), the Sloan Digital Sky Survey (center), and the S2 data (right), all centered on the same prominent large-scale structure feature: the Sloan Great Wall. The S2 data at right shows a slightly distorted projection of the actual Sloan Great Wall feature.

Credit : T. Sawala, arXiv:2607.01172, 2026

However, we have to be careful: are these actual departures from the standard picture of isotropy, or are they an artifact of how the data were selected, calibrated, and handled?

Take a look at the three graphs in the image above. In the center is the Sloan Great Wall, mapped out from the Sloan Digital Sky Survey in black, with additional (blue) data points showing an overlay of DESI data. On the left is the sources above a certain brightness identified from DESI data, showing a more comprehensive mapping of the Sloan Great Wall and the region around it, which also goes to much greater distances and sensitivities.

Then, over on the right, you can see the main “bright galaxy sample” chosen in the Sylos Labini & Galoppo paper, which appears distorted compared to the regular DESI data. In fact, that distortion is not only real, but important: it’s a special type of distortion occurring along the line-of-sight known as redshift space distortions . In our actual Universe, you have large overdensities: massive clumps of matter. Within the largest of these clumps of matter, galaxy clusters and great galactic walls form. Individual galaxies move around inside of these objects omnidirectionally, but there’s a catch.

Some galaxies appear to be approaching us, as they accelerate towards the cluster’s center from behind, giving them a relative blueshift and causing us, because of that additional shift, to place them at a shorter distance (corresponding to a lower redshift).

Other galaxies appear to be receding from us, as they accelerate towards the cluster’s center from in front, giving them a relative redshift and causing us, because of that additional shift, to place them at a greater distance (corresponding to a greater redshift).

That’s why the picture on the right has a distorted, smeared-out wall: because of these redshift space distortions. Poetically, these are also called “Fingers of God,” because they point at you: the zero-redshift observer.

FOGs, or Fingers of God, are known to appear in redshift space. Because galaxies in clusters can get extra redshifts or blueshifts due to the gravitational influence of its surrounding masses, those galaxy positions that we infer from redshift will be distorted along our line-of-sight, leading to the Fingers of God effect. When we perform our corrections and move from redshift space (left) to real space (right), the FOGs disappear.

Credit : M. Tegmark et al., Astrophysical Journal, 2004

If you don’t properly account for these redshift-space distortions, you’re going to see a boost in power on large cosmic scales compared to what you expect to observe, while simultaneously seeing a suppression in power on small cosmic scales compared to what you’d expect.

The main result of the Sylos Labini & Galoppo paper is exactly consistent with that scenario: they report seeing an excess of power on very large cosmic scales, and on gigaparsec (~3+ billion light-year) scales in particular, which are larger scales than should be possible to have an excess of power on in our (presumably) isotropic, homogeneous Universe.

However, another independent analysis was done: by Till Sawala , who you might recognize as the lead author of a recent paper that showed that the long-expected Milky Way-Andromeda merger may not occur in 4-5 billion years , but rather likely requires much longer timescales. By properly accounting for these redshift-space distortions, Sawala was able to show that this claimed “excess power” effect on large cosmic scales goes away entirely , and that instead of violating the cosmological principle, the observed structures in the DESI data actually are consistent with the ones expected in our Universe, and are entirely consistent with what you get when you run your own cosmological simulations.

The big discrepancy between the Sylos Labini & Galoppo study (dot-dash line, using DESI data) and the analysis of Sawala (dashed line, also using DESI data) is caused by the former team failing to comprehensively account for redshift space distortions, and instead using fiducial coordinates. When the correct comoving distance scale is used, the apparent anomaly vanishes, with the dashed curve matching simulations (such as the pink line) exquisitely.

Credit : T. Sawala, arXiv:2607.01172, 2026

Sylos Labini, perhaps foreseeing that his conclusions would be refuted in short order, hedged his bets when speaking to journalists , stating:

“Ultimately, the question is not whether our paper is right or wrong. The question is whether nature is telling us something new about the universe on the largest scales.”

That is wrong. This paper made an extraordinary claim: one that is factually wrong. The discrepancy claimed in Sylos Labini’s paper is, as Sawala showed, “plausibly explained by a distance error inherited from the construction of the S2 coordinates… [which] treat luminosity distances… as if they were comoving distances.” You cannot build a case for investigating the Universe for large-scale anomalies atop a flawed foundation like this; the entire line of thinking simply falls apart.

The question of whether nature is telling us something new about the Universe on the largest scales of all is one that we should keep in the back of our minds only, as the full suite of data has been precisely what has led us to accept the consensus picture of the Universe in the first place. We must keep an open mind to the possibility that there may be something new that future data will reveal, but we absolutely must not expect such a revolution. It is that kind of flawed thinking that leads us to “see anomalies” where none exist: because we’re allowing motivated reasoning or wishful thinking to influence what should be sober, scrupulous, merit-based scientific work.

This is a big danger associated with conducting cutting-edge science: that we’re so focused on finding something that runs counter to the mainstream that we elevate fundamentally flawed research that claims to find it. This then gets amplified in the popular media, eroding the trust that scientists have worked so hard — and so scrupulously — for so long to earn. It further underscores why we need expertise and careful, mistake-free work, rather than sensationalism and hope for a revolution, to be our scientific guides. Without them, we risk falling into the trap that Sylos Labini advocates for: ignoring the question of what’s right or wrong solely on the basis of its scientific merits. If we wish to do science well, there is no other way.

This article Famous study in error: the Universe isn’t anisotropic is featured on Big Think .