The discovery of an extremely metal-poor galaxy, observed thanks to the James Webb Space Telescope, opens a fissure in the traditional cosmological narrative and invites us to think of the Universe not as a neat succession of closed stages, but as a rough, non-homogeneous process filled with persistent remnants.
For decades, the history of the Universe has been told as an orderly sequence: first a young, simple, primitive cosmos; then a mature, enriched and complex one. A clean, pedagogical, efficient narrative. Too efficient.
A new study recently published in The Astrophysical Journal Letters, entitled “A Metal-Free Galaxy at z = 3.19? Evidence of Late Population III Star Formation at Cosmic Noon”, has introduced a deep discomfort into that story. Astronomers have identified an extremely metal-poor galaxy —that is, almost devoid of heavy chemical elements— at a time when, according to standard models, such systems should no longer exist. The galaxy, named CR3, does not belong to the infant universe, but to a cosmos that had already passed through its adolescence.
The finding does not directly contradict current cosmology. It does something more unsettling: it reveals its texture.
A galaxy that arrives late to its own history
In astronomy, “metals” are not jewels or ingots, but any element heavier than helium: carbon, oxygen, iron. These elements were not born with the Big Bang. They were forged inside stars and dispersed into space when those stars died in violent explosions.
For this reason, the presence of metals functions as a chemical clock. The more metals a system contains, the more stellar history it has lived through.
Galaxy CR3 presents a problem: it seems to have lived through almost none. Its spectrum shows strong signatures of hydrogen and helium, but lacks —or almost lacks— the lines that would reveal the presence of metals. Moreover, the radiation it emits is extremely hard, compatible with very young and massive stars, similar to the hypothetical Population III stars, the first to form in the Universe.
The crucial detail is temporal: CR3 is observed at a redshift of z ≈ 3.19, meaning that its light was emitted when the Universe was around two billion years old. Redshift —the stretching of light due to the expansion of space— indicates not only distance, but also age: the higher the redshift, the further back in time we are looking. In this case, we are observing a universe that had already passed the great transition of reionization and that, according to the standard narrative, should have been chemically mature.
It should not be there.
And yet, it is.
Why we can see it now
This discovery would not have been possible without the James Webb Space Telescope (JWST). Unlike previous instruments, JWST is designed to observe in the infrared, precisely where the light from very distant objects arrives today, having been originally emitted in ultraviolet or visible wavelengths and stretched by the expansion of the Universe.
Webb does not “see better” in a classical sense: it sees where we could not look before. It is specifically specialized in reading spectra displaced by redshift, allowing scientists to identify the chemical composition and stellar age of remote galaxies like CR3. In other words, the right instrument arrived just as the Universe began to show its wrinkles.
The problem is not the model. It is the scale
To understand why this discovery matters so much, we must abandon a common confusion: the belief that cosmological models describe the Universe in all its details.
They do not.
They never have.
Models work on average. From afar. Just as the Moon, seen from Earth, appears as a smooth and perfect disk. Only when one gets closer do craters, fractures and layered structures become visible. The Moon did not change. The scale of observation did.
Cosmology has long observed the Universe as that distant disk: clean stages, clear transitions, smooth edges. The article does not break the disk. It reveals its craters.
A cosmic “cototo” (a lump, a bump)
The best way to understand galaxy CR3 is not as an anomaly, but as a remnant. A “cototo”, using a colloquial but precise word: a persistent irregularity of the process.
The expansion of the Universe was never perfectly homogeneous. Chemical mixing was never instantaneous. Supernovas enrich specific regions; others remain isolated, protected by low density, geometry, or simple contingency.
Just as medieval technologies coexist with satellites on Earth, very different chemical histories can coexist in the cosmos.
This finding suggests that the Universe does not neatly erase its past. It overlays it. It drags it along. It allows it to survive in unexpected corners.
When “being” is no longer enough
This discovery forces a deeper conceptual correction: abandoning the idea that things simply are.
In contemporary physics, matter is not a solid substance. It is a set of interacting fields. Particles are events. Galaxies are dynamic equilibria. Even space-time is not a fixed stage, but a geometry that evolves.
Nothing “is” in a static sense.
Everything is being.
Galaxy CR3 does not belong to a closed stage of the past. It is being a different history within a larger process. Like a wrinkle on a human face: not an error of time, but its materialized trace.
The universe does not advance. It resists.
If this result is confirmed, it does not change the origin of the Universe. It changes the way we read its history.
The cosmos does not advance like a clean arrow toward complexity. It advances by leaving behind remnants, exceptions, memories. Like every real process.
The Universe is not a straight line.
It is a rough surface.
And only now are we learning how to touch it. It is fascinating.
Note:
The study referred to in this article is “A Metal-Free Galaxy at z = 3.19? Evidence of Late Population III Star Formation at Cosmic Noon”, by Sijia Cai, Mingyu Li, Zheng Cai, Yunjing Wu, Fujiang Yu, Mark Dickinson, Fengwu Sun, Xiaohui Fan, Ben Wang, Fergus Cullen, Fuyan Bian, Xiaojing Lin and Jiaqi Zou, published in The Astrophysical Journal Letters (volume 993, issue L52, November 2025). The full article can be consulted in the arXiv preprint repository: 2507.17820 or directly on the journal’s website.
