A team of astronomers has used the James Webb Space Telescope to chart 164,000 galaxies across billions of light-years, producing the most detailed three-dimensional map ever assembled of the vast, mostly invisible scaffolding that holds the universe together. The result, published in May 2026 and built on the COSMOS-Web survey, reaches back to a time when the cosmos was less than a billion years old and reveals the filaments of matter and dark matter along which galaxies formed and grew.
“We’re seeing the skeleton of the universe at a level of detail that simply wasn’t possible before Webb,” said Caitlin Casey, an astrophysicist at the University of Texas at Austin and a principal investigator of the COSMOS-Web program. The map, she noted, traces structure across a period when the first substantial galaxy clusters were just beginning to take shape.
The cosmic web, explained
On the largest scales, matter in the universe is not scattered randomly. It is organized into a network sometimes called the cosmic web: long filaments of gas and dark matter that stretch hundreds of millions of light-years, intersecting at dense nodes where galaxy clusters sit, and separated by enormous near-empty voids. Think of it as the universe’s connective tissue. Galaxies form along the filaments and pile up at the intersections, pulled there by gravity over billions of years.
Mapping this web is difficult because most of its mass is dark matter, which emits no light. Until now, the sharpest maps covered relatively nearby stretches of the cosmos or relied on instruments that could not see the faintest, most distant galaxies. Webb changed that equation with its powerful infrared cameras, which can detect light from galaxies so far away that their signals have been stretched beyond what optical telescopes can capture.
How the map was built
The new map draws on deep infrared images taken by Webb’s Near Infrared Camera (NIRCam) across a patch of sky in the COSMOS field, a well-studied region about 0.6 square degrees wide. The underlying galaxy catalog, called COSMOS2025, combines data from Webb, the Hubble Space Telescope, and ground-based observatories to provide positions, brightnesses, shapes, and distance estimates for more than 700,000 galaxies.
From that larger catalog, the team selected roughly 164,000 galaxies with the most reliable distance measurements, known as photometric redshifts, which are derived from multiband imaging rather than from individual spectra. By analyzing where those galaxies sit in three dimensions, the researchers reconstructed the filaments, voids, and dense nodes of the cosmic web out to a redshift of about 7, corresponding to roughly 800 million years after the Big Bang.
A companion study, published in Nature Astronomy, went further by using a technique called weak gravitational lensing. Gravity from foreground mass subtly warps the shapes of background galaxies, and by measuring those tiny distortions across millions of pixels, the team charted the distribution of total mass, including the dark matter that makes up the bulk of it. The resulting dark matter map is the highest-resolution chart of its kind ever produced at these cosmic distances.
As NASA noted in its summary of the findings, Webb’s infrared sensitivity detects fainter and more distant galaxies than Hubble could, filling in structure at epochs that were previously invisible. The wide survey area was chosen specifically to reduce a statistical problem called cosmic variance, the risk that a small patch of sky might not be representative of the universe as a whole.
Early signs of a problem with existing models
The map is not just a prettier picture. Early indications from the COSMOS-Web data suggest that matter clumped together more densely at high redshifts than some leading cosmological simulations predict. If that pattern survives further scrutiny, it would mean that the interplay between dark matter halos and ordinary gas was more efficient in the early universe than standard models assume, or that simulations are not accurately capturing processes like star formation feedback and the growth of supermassive black holes.
That possibility has drawn attention because it echoes a broader pattern. Since Webb began science operations in 2022, several studies have found unexpectedly massive or mature galaxies at very early cosmic times, raising questions about whether the standard timeline of galaxy assembly needs revision. The COSMOS-Web structure map adds a new dimension to that debate by showing not just individual galaxies but the large-scale environment in which they sit.
Still, confirming any real tension with theory will require cross-checks against independent surveys and comparisons to multiple simulation suites, not just a single benchmark model.
What the map cannot yet show
For all its detail, the map has important limitations. The 164,000-galaxy sample relies on photometric redshifts rather than spectroscopic confirmations, which are far more precise but enormously time-consuming to collect for this many objects. Photometric estimates can misplace some galaxies along the line of sight, effectively smearing out structures in three dimensions. How much that blurring affects the reconstructed filaments and voids has not been fully quantified in published analyses.
The weak lensing measurements carry their own challenges. Extracting the faint distortion signal requires extremely precise galaxy shape measurements, and systematic effects from the telescope’s optics or from the algorithms used to estimate shapes can mimic or mask real signals. The Nature Astronomy paper reports extensive consistency checks, but independent replication using alternative shape-measurement pipelines has not yet appeared in separate publications.
There is also a brightness floor. Even Webb can only detect galaxies above certain luminosity thresholds, and those limits tighten at earlier cosmic times. The current map therefore traces the cosmic web using relatively massive, luminous systems. How faithfully that skeleton represents the full distribution of matter, including faint dwarf galaxies and diffuse gas, remains an open question.
What comes next
Several efforts are already underway that could sharpen or challenge the COSMOS-Web results. The European Space Agency’s Euclid mission is conducting its own wide-field weak lensing survey at optical and near-infrared wavelengths. Joint analyses between Euclid’s broad coverage and Webb’s depth could reveal whether the dense high-redshift clustering seen in the COSMOS field is a local fluctuation or a genuine cosmological signal. As of June 2026, no detailed joint analysis timeline has been announced by either collaboration.
Spectroscopic follow-up will also be critical. Programs using Webb’s own spectrographs and ground-based instruments like the Keck telescopes and the forthcoming Extremely Large Telescope could confirm the photometric distances for a subset of the 164,000 galaxies, testing whether the three-dimensional structure holds up under more precise measurement.
For now, the COSMOS-Web map stands as the deepest and most detailed look at the architecture of the early universe. Whether it ultimately sharpens the existing picture of cosmic evolution or forces astronomers to rethink how structure formed in the first billion years, the data are already reshaping the questions the field is asking.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
