The Cosmic Web Just Broke the Big Bang Model

Look out into the universe and matter does not appear evenly spread. Stars gather into galaxies, galaxies gather into clusters, clusters connect into vast filaments, and those filaments surround enormous cosmic voids. This is the cosmic web, a structure so large that it only becomes visible when we map the universe on the biggest scales. But in the standard Big Bang model this cannot simply continue forever. So what would happen if a new survey found that it did? This is what a paper published in Nature Now claims. And if correct, the implications for our understanding of the universe could be profound. At first it might seem strange that there should be any limit to how far a structure can extend through the cosmic web. After all, if galaxies can gather into clusters, and clusters can connect through filaments, why should those filaments not continue onto ever larger scales? The answer lies in how the Big Bang model builds structure. I've covered some of the problems with this before, because structure formation in Lambda CDM is not as simple as saying gravity did it. In the standard picture, the universe begins in a hot, dense, rapidly expanding state. It doesn't begin with galaxies, clusters, and filaments already in place. It begins very close to smooth with only tiny differences in density. And that creates a problem. In the early hot phase, ordinary matter is tightly coupled to radiation. Photons are constantly scattering from charged particles. Radiation pressure keeps pushing back. So ordinary matter can't simply collapse into galaxies right away. The universe has to expand and cool first. Only then can matter and radiation decouple, allowing ordinary matter to fall more freely under gravity. That is why Lambda CDM needs dark matter to do much of the early work. Dark matter doesn't couple to radiation in the same way. So while ordinary matter is still being held back by the radiation field, dark matter can already begin gathering under gravity. Density differences grow into deeper gravitational wells. Those wells become the hidden framework of structure. Then, once the universe has expanded and cooled enough, ordinary matter can fall into those dark matter wells. Over time, the gas and plasma gathers, cools, and forms galaxies. Those galaxies collect into groups and clusters, while the surrounding matter stretches into filaments and leaves vast underdense voids between them. That is how Lambda CDM builds the cosmic web. But this process is not unlimited. The model begins with tiny initial fluctuations, gravity needs time to amplify them, and expansion does two things at once. It helps the universe cool, making structure formation possible. But it also limits how much structure can grow, and how far separate regions can become physically connected. So from the model's point of view, the cosmic web should have a coherent scale. Below that scale, structure can form. But beyond that scale, the pattern should start to average out. The universe should become statistically smoother. So the real question is how far that structure should remain coherent before the cosmic web gives way to the large-scale smoothness the model expects. And this is what the new paper is testing. Not by claiming to have found one enormous object in space, or a single giant filament, but asking a more basic question. When galaxies are separated by a certain distance, do they point in random directions? Or do they still line up as part of a larger structure? Imagine taking two galaxies and drawing a line between them. That line has a length, but it also has a direction. Now do that not once, but millions of times across a huge galaxy survey. If the universe has become statistically random on those scales, those lines should point in random directions. Some left to right, some up and down, some diagonal. No direction should keep winning. But if galaxies are still sitting inside larger filaments, walls, sheets, or connected structures, then many of those lines will point in the same direction. The pattern has not washed out. And this is what the paper claims to detect. The authors find that galaxy pairs do not become randomly oriented at the scale expected. Instead, the real galaxy data shows persistent alignment. Preferred directions that should have washed out. And when they compare this to LambdaCDM simulations, built with the same geometry and the same kinds of survey slice, the simulations do not show the same persistent pattern. In the model, the direction fades. In the data, they remain. That is why this is more serious than simply saying we found a big thing. A single, large object could be unusual. A statistical pattern across the galaxy distribution suggests something deeper. It suggests the cosmic web itself may be organised on scales where the standard model expects that organisation to disappear. So why not just adjust the model? Why not make the early universe a little less smooth? Why not give gravity a little more help? The difficulty is that LambdaCDM is tightly constrained. It's not free to create any structure it wants. Make the initial fluctuations larger and you run into the microwave background. Make the growth faster and you affect galaxy surveys and the age of the universe. Allow large-scale structures to keep going without averaging away and you're no longer making a small adjustment. You're changing the background assumptions of the model. You can always add new physics, self-interacting dark matter, modified gravity, back reactions. But each one of these has to do more than solve one problem. It has to preserve everything else the model is usually taken to explain. That's why this paper matters. This is bigger than a small correction to the cosmic web. It's asking whether the assumptions used to build the cosmic web are strong enough to explain the universe we actually observe. Because if coherent structure really continues to gigaparsec scales, then the problem is not just the web is larger than expected, the problem is that the model may be starting from the wrong kind of universe. One anomaly is easy to ignore. One awkward measurement can be treated as a statistical accident, a selection effect, or a method that asks the wrong question. It becomes much harder when different analyses keep pointing to the same problem. And that is why the results matter In that earlier paper, the question was different. It was not about whether galaxy pairs point in preferred directions. It was about density. As you look on larger and larger scales, does the average density of galaxies settle down? Does the universe eventually become a smooth background, with local structures sitting on top of it? Using a different way of asking that question, Labini argued that the galaxy distributions did not converge towards the uniform density. At the time, you could push that aside. Maybe the result depended on the statistics being used. Maybe standard cosmology asked the question differently. Maybe it was interesting but not decisive. But now this new paper adds a second layer. This time the question is not density, it is direction. Do the structures lose their preferred alignments? Does the organisation wash out? And again, the answer appears to be no. So the two papers are not making exactly the same claim. That is what makes them powerful together. Different tests, same tension, one points to density, the other points to direction. But both suggest that, at the larger scale we can map, the universe is not averaging away as cleanly as LambdaCDM expects. There is also another possible layer here. Because if structure remains coherent on these scales, it may not only affect where galaxies are, it may affect how galaxies move. Modern cosmology doesn't just map positions. It uses redshift, distance, motion and reference frames together. It assumes that, on large enough scales, local motion and bulk flow can be separated from the smooth Hubble expansion. But if the structure itself has not averaged away, then perhaps that motion has not averaged away either. That begins to connect this paper to other questions I've covered before. Bulk flows, cosmic dipoles, possible directional effects in the Hubble flow and the possibility that some cosmological tensions are not really separate problems, but different ways of seeing the same large scale failure to average out. I'm going to unpack that more carefully in a Patreon reflection, because it's too big to rush through here. But for this video the point is simpler. The new paper may not be an isolated anomaly. It may be part of a wider pattern where density, direction and motion all keep pointing back to the same problem. And that leaves an unavoidable question. Maybe the standard model is not the final answer. Maybe the universe is not built from tiny early fluctuations that slowly grow into the structure and then fade into a smooth background. That opens the door to a different kind of cosmology. A fractal universe, a plasma universe, a model where structure is more fundamental, more persistent and more connected across scale. At first that sounds simple. If another model naturally expects structure this large, then why not use that model instead? But that only opens the harder problem. Because lambda CDM, for all its problems, didn't become the standard model by explaining one observation. It became the standard model because it appeared to connect many different observations into one framework. So if a new result cracks open the framework, the next question is not simply which alternative can explain this one crack. The next question is much harder. If lambda CDM is under pressure, what evidence must any serious alternative be able to explain?