Nobel Prize Winner Warns “This Isn’t Our Universe” — James Webb Found Something Strange

Ajouté : 2026-05-10

[00:00:00]We have misunderstood the universe.

[00:00:02]Those are not my words. Those are the words of Adam Ree, a Nobel Prizewinning physicist at Johns Hopkins University. A man who has spent his career measuring how fast the cosmos is flying apart. And he is not trying to be dramatic. He is reading his own data and telling us plainly that something in our model of reality does not add up. For nearly a century, cosmology has rested on a single elegant framework. We call it the standard model, a set of equations and assumptions that describes everything from the first fraction of a second after the big bang. It has worked beautifully for decades, explaining observation after observation with an almost suspicious level of precision.

[00:00:39]And then the James Webb Space Telescope turned on. What Webb found did not confirm our understanding. It made things more complicated. In some cases, it directly contradicted it. And the deeper we have looked with this telescope, the wider the cracks have become. The trouble starts with a single number. It is called the Hubble constant. And it measures the rate at which the universe is expanding right now. Not in the past, not in some theoretical future. Right now, today, at this very moment. You would think a number like that would be easy to pin down. Two methods, two teams, same universe, same answer. That is how science is supposed to work. But it did not. And this is not the first time. The Hubble constant has been causing argument since the 1920s when Edwin Hubble himself first calculated it. His original estimate was so far off that it made the universe appear younger than the Earth. For decades after that, two camps of astronomers fought bitterly over whether the true value was closer to 50 or 100. A disagreement so fierce it split the field in half. By the early 2000s, the number had finally converged.

[00:01:42]Everyone agreed it was somewhere around 70, give or take. The war was over. The measurement problem was solved. Except it was not. Because as instruments got more precise, a new split emerged.

[00:01:52]Smaller than the old one, but far more stubborn. The first method looks backward, way backward. The European Space Ay's Plank satellite spent about 4 and 1/2 years mapping the cosmic microwave background. The faint afterlow of radiation left over from when the universe was about 380,000 years old.

[00:02:09]From that map, physicists extracted the conditions of the early universe, plugged them into the standard model, and ran the equations forward nearly 14 billion years to predict how fast the universe should be expanding today. The answer they got was about 67 km per mega parseek. A mega parseek, by the way, is 3.26 million lighty years. So for every 3.26 million lighty years you look outward, galaxies are moving away from us 67 km faster. a clean number, a tiny margin of error, confidence level through the roof. The second method is more direct. Instead of modeling the past and projecting forward, it measures the present. Adam re and his team called SH0ES use a technique known as the cosmic distance ladder. They start by measuring the distance to nearby pulsating stars called cified variables.

[00:02:56]These are stars whose brightness fluctuates at a rate directly tied to how luminous they actually are. Measure the pulse, calculate the true brightness, compare it to how bright the star looks from Earth, and you have got a distance. From there, they step outward to galaxies containing both teids and a specific type of exploding star called a type IA supernova. These supernova are useful because they all explode with roughly the same peak brightness. So once you have calibrated them with sephieds in the same galaxy, you can spot them in galaxies hundreds of millions of light years away and calculate how far those galaxies are.

[00:03:29]Combine that with how fast they are receding and you get the Hubble constant. Ree's team got 73, not 67. 73 6 km per mega parseek faster than the standard model predicts. That might sound like a rounding error. It is not.

[00:03:44]In a field where measurements are routinely precise to within 1%, a gap of 8 or 9% between two independent methods is enormous. It is the difference between a universe that is about 13.8 billion years old and one that could be more than a billion years younger. It changes the size of the observable universe. It changes the rate at which galaxies are separating. It changes potentially the nature of the dark energy that has been pushing everything apart since the late 1990s. Think about what that means for a second. If the universe is expanding faster than our best model predicts, then everything downstream of that prediction shifts.

[00:04:17]The age of the universe gets younger.

[00:04:19]The distances between galaxy clusters change. The timeline for when the first stars ignited, when the first heavy elements were forged inside dying suns, when the conditions for planets and chemistry and eventually life became possible, all of it shifts. You are not just adjusting a dial on a dashboard.

[00:04:36]You are redrawing the entire history of everything that has ever existed. For a while, scientists assumed someone had made a mistake. The most popular theory was stellar crowding. Sefeed variables tend to live in the dense dusty discs of younger galaxies. When Hubble looked at them, its cameras sometimes blended the light of nearby stars together, making the seaweeds appear brighter than they truly were. Brighter seafeds mean shorter calculated distances. Shorter distances mean a higher Hubble constant.

[00:05:06]If you could clean up that blending, the argument went, the tension would relax.

[00:05:10]So they repointed the James Webb Space Telescope at the same stars. Web orbits the sun about 1.5 million kilometers from Earth, far beyond the interference of our planet's atmosphere and thermal radiation. Its 6.5 meter gold-plated mirror collects infrared light with a sensitivity that makes Hubble look like a pair of reading glasses. If any telescope could settle this, it was web.

[00:05:30]The expectation among many in the field was that web would gently deflate the tension, that its sharper vision would reveal the blending, correct the distances, and bring the Hubble constant back into line with the standard model.

[00:05:43]Some astronomers were already drafting the obituary for the crisis before the data came in. Web's infrared cameras can cut through dust and crowding with a precision Hubble never had. If blending was the problem, Web would show it. The seafs would look dimmer through Web's eyes. The distances would stretch and the Hubble constant would drop back towards 67. Crisis averted. That is not what happened. Webb confirmed Hubble's numbers. Not approximately, but almost exactly. The Sephieds looked the same.

[00:06:10]The distances held. The expansion rate stayed at 73. In early 2024, the Shadzer team published a paper in the astrophysical journal letters announcing that they had now spanned the full range of Hubble's Sephiide observations with web extending all the way out to a galaxy called NGC 5468 about 130 million light years from Earth, a thousand sephides across five host galaxies and eight supernovi. The conclusion was clear. measurement error could be ruled out with very high confidence. Ree put it simply, with the errors eliminated, what is left is the possibility that we have misunderstood the universe. At a 2019 conference at the Cavi Institute for Theoretical Physics in California, re asked David Gross, a Nobel laureate in particle physics, whether the field should start calling this discrepancy a problem. Gross corrected him. He said they should not call it a tension and they should not call it a problem. They should call it a crisis. By December 2024, the crisis had deepened. RE's team published their largest web study yet.

[00:07:15]Using data collected across the telescope's first two years of operation, they employed three separate methods to measure distances to supernova host galaxies, cross-checking seeds against carbon stars, and a technique called the tip of the red giant branch. All three methods agreed with Hubble's original measurements. All three pointed to a universe expanding faster than the standard model allows.

[00:07:39]So either the model is wrong or something is hiding inside the physics that we have not accounted for. Not everyone agrees the tension is even real. Wendy Freriedman, an astronomer at the University of Chicago, has spent years building an independent distance ladder that relies less heavily on sea feeds. Her team uses additional types of stars, red giants, and a class called Jagbe stars that are found in less crowded regions of galaxies and are less likely to be affected by dust and blending. When she submitted her own web analysis in August 2024, her three methods gave different answers. The red giant and jagged B methods landed around 68 to 70, consistent with the standard model. Her seafide measurement came in higher, near 72, but with larger uncertainties. Her overall combined result was about 70 sitting in a kind of no man's land between the two camps.

[00:08:26]Saul Pearlmutter, a Nobel Prize-winning cosmologist at UC Berkeley, who reviewed Freriedman's data before its release, noted that the results suggest there may be a tension within the star-based measurements themselves before you even compare them to the cosmic microwave background. That is a different kind of problem. It means the distance ladder, the tool astronomers have relied on for a century to map the nearby universe, might have unresolved issues at its foundation that no single telescope can fix. Re pushed back. He argued that Freriedman's team had used a small and potentially unrepresentative subset of supernovi in their analysis, which could bias the results downward. The disagreement between the two camps has become one of the most closely watched rivalries in modern physics. Two teams, both brilliant, both careful, both looking at the same sky, arriving at different conclusions about what the sky is telling them. One possibility that keeps surfacing is called early dark energy. A hypothetical burst of extra expansion that occurred in the first few hundred thousand years after the Big Bang, then vanished. If something like that happened, it would change the conditions in the early universe just enough to shift the plank prediction upward, closer to 73, without breaking the other things the standard model gets right. Mark Cameianowski, a cosmologist at John's Hopkins who has helped develop this idea, has compared it to the universe receiving an unexpected kick right after it was born. A nudge that would have rippled forward through 14 billion years of expansion, slightly altering the rate we measure today.

[00:09:51]Nobody knows if early dark energy is real. No one has detected it. But it is one of the few ideas that could solve the Hubble tension without demolishing everything else that works. the tension would be enough on its own. But web delivered a second blow almost as soon as the telescope started observing. It found galaxies in the early universe that had no business being there. Not small, dim infant galaxies, which is what the standard model predicted at those distances, but massive ones.

[00:10:20]Galaxies seen as they existed just 500 to 700 million years after the Big Bang.

[00:10:25]Some of them rivaling the Milky Way in mass. A galaxy that took our own neighborhood 13 billion years to build had apparently assembled itself in under a billion. Mike Boland Colchin, an astronomer at the University of Texas at Austin, ran the numbers in a study published in Nature Astronomy. He showed that six of the earliest and most massive galaxy candidates observed by web pushed right up against the absolute limit of what the standard model allows to form galaxies that heavy that fast.

[00:10:55]Nearly every available atom in their surrounding dark matter halos would have had to convert into stars. In a normal universe, that conversion rate sits around 10%. 90% of the gas just never gets around to forming anything. it is too hot or it gets blown out by radiation or it simply drifts. These galaxies seem to be running at 100% efficiency like a factory that somehow turns every scrap of raw material into finished product with zero waste.

[00:11:21]Physics does not work that way. Not in any model anyone had built. The initial reaction was frankly alarm. If galaxies this massive really form this quickly, then either our models of galaxy formation are deeply incomplete or the standard model of cosmology itself needs revision. Some physicists proposed that new particles or forces might have existed in the early universe, accelerating star formation beyond anything current models predict. Then came a correction. Katherine Chwowski, a graduate student also at the University of Texas, led a study published in mid 2024 that re-examined the data. She found that many of the apparently over massive galaxies were being fooled by their own black holes. These galaxies, nicknamed little red dots because of their color and compact size, hosted black holes that were rapidly devouring gas. The friction from all that infalling material heated it to extreme temperatures, and that heat radiated outward as light, lots of it, blending with the starlight of the galaxy itself.

[00:12:21]From billions of light years away through Web's cameras, a galaxy with a hungry black hole at its center looks almost identical to a galaxy with 10 times more stars. The telescope sees brightness, and brightness is usually a reliable proxy for mass. Usually, in these cases, the black hole was doing the work of billions of suns, inflating the galaxy's apparent size like a flashlight taped to a candle. Once those little red dots were removed from the sample. The remaining galaxies fit within the standard model's predictions.

[00:12:53]Steven Finkelstein, Churovsky's adviser, was blunt about it. There is no crisis in terms of the standard cosmological model, he said. Not from the galaxies at least. But the story did not end there cleanly. Even after removing the little red dots, roughly twice as many massive galaxies remained as the standard model expects. Not enough to break the model, but enough to bend it. Chorovski suggested that stars may have formed faster in the early universe than they do today. That the denser conditions shortly after the Big Bang made it harder for gas to escape during star formation, allowing the process to accelerate. A reasonable explanation, but one that still requires adjustments to how we think galaxies grew up. And then in early 2025, Webb found something else. A massive spiral galaxy quickly nicknamed the big wheel that existed within the first two billion years of the universe. Spiral structure takes time to develop. It requires relatively calm, orderly growth, not the chaotic merging you would expect in a young universe. Femia Nanayyakura, one of the astronomers who discovered it at Swinburn University of Technology, said the galaxy either assembled its mass in an unusually neat way or it formed most of its stars in place without the violent collisions that typically build large galaxies. Either explanation challenges conventional thinking.

[00:14:10]Meanwhile, a separate team at Cambridge used web to study over 250 galaxies from that same era and found the opposite pattern. Most of them were chaotic, turbulent, clumpy systems that had not settled into smooth rotating discs. So, the early universe contained both messy adolescent galaxies and at least one remarkably mature spiral coexisting in the same epoch. That is not a contradiction the standard model cannot handle gracefully. But none of this addresses the elongated galaxies. A study published in Nature Astronomy in late 2025 led by Alvaro Poso and including researchers from MIT, Harvard, and Taipei found that many young galaxies observed by Web appear strikingly stretched out like cosmic cigars. These prolate shapes do not match predictions from the cold dark matter model which is the backbone of standard cosmology. They do however match predictions from alternative models like warm dark matter and wave dark matter in which the filaments of the large-scale structure of the universe are smoother allowing gas and stars to flow along them in elongated streams. That distinction matters more than it sounds. Cold dark matter is not just one ingredient in the standard model. It is the scaffolding. It determines where galaxies form, how they cluster, how the large scale structure of the universe looks on the biggest scales. If dark matter particles are lighter or behave differently than the cold model assumes, the ripples from that change would propagate through almost every prediction cosmology makes.

[00:15:37]It would alter the mass of galaxy clusters. It would shift the expected distribution of dwarf galaxies around larger ones. A distribution that has already been a source of tension for years. It would change how gravitational lensing bends light around massive objects. You cannot swap out the foundation of a building and expect the walls to stay where they are. So where does this leave us? The standard model is not dead. Not even close. It still explains the cosmic microwave background with extraordinary precision. It predicted the abundances of hydrogen and helium in the universe. It accounts for the clustering of galaxies across billions of light years. No competing model can match its track record. But it is developing fractures that were not there a decade ago. The Hubble tension refuses to go away. The early universe is producing galaxies that are bigger, more structured, and stranger than expected. The nature of dark matter, which makes up roughly 27% of everything that exists, may not be what we have been modeling for the past 25 years. And dark energy, the 68% of the universe responsible for its accelerating expansion, remains a placeholder name for something we cannot identify, measure directly, or explain. There is even a second, quieter tension that most people have not heard about. It is called the S8 tension and it concerns how clumpy the universe is. The standard model predicts a certain amount of clustering in the distribution of matter across space. When astronomers measure that clustering directly, they consistently find slightly less of it than the model expects. Adam Ree has called it the little sibling of the Hubble tension. Less dramatic, but worth watching. If both tensions turn out to be real, they may be symptoms of the same underlying problem. Two cracks in the same wall. Adam Ree put it carefully in December 2024, as he often does. The discrepancy between the observed expansion rate and the predictions of the standard model suggests that our understanding of the universe may be incomplete. Two NASA flagship telescopes now confirm each other's findings. We have to take this problem seriously.

[00:17:35]What is remarkable about this moment in cosmology is that the anxiety is productive. Nobody is panicking. The prevailing mood is closer to the feeling you get when you have been assembling a jigsaw puzzle for years, confident you had the picture right, and then you find a piece that does not fit anywhere. It does not mean the puzzle is wrong. It means the picture on the box might not be the whole picture. New instruments are coming. NASA's Nancy Grace Roman Space Telescope will conduct wide field surveys designed specifically to study dark energy. The European Space Ay's Uklid mission is already mapping the geometry of the universe on a scale no telescope has attempted before. Future data releases from the Gaia Space Telescope will let astronomers calibrate cified distances with geometric precision, removing another potential source of error. And the Adakama Cosmology Telescope has already produced the most detailed groundbased map of the cosmic microwave background ever made, slightly nudging the plank estimate upward, but not far enough to close the gap. We are living inside a question.

[00:18:35]The most precisely measured universe in history is telling us two different things at the same time. And both of them appear to be correct. Something in the space between the beginning and now in those 14 billion years we have never directly observed is doing something we do not understand. Some ingredient we have not identified, some process we have not modeled, some property of space itself we have not imagined. Cosmology has been here before. In the 1990s, the field nearly collapsed under the weight of a different contradiction.

[00:19:03]Measurements suggested the universe was younger than its oldest stars. That made no sense. The stars could not be older than the thing they lived inside. Then came the discovery that the expansion of the universe was accelerating, driven by dark energy, and the whole picture snapped into focus. What looked like a fatal flaw turned out to be the shadow of something enormous that nobody had seen yet. That is the thing about crises in physics. They feel like endings. They almost always turn out to be doors. If you could somehow stand outside the universe and watch this moment, you would see a species on a small rocky planet orbiting an unremarkable star in the suburbs of a midsized galaxy, using mirrors and math to argue about whether reality is expanding 6% faster than it should be and getting genuinely upset about it. That is either absurd or magnificent. And I think it might be both. It is not a failure. Honestly, it is the opposite. This is what it looks like when a science reaches the boundary of what it knows and keeps pushing. The universe is not broken. Our map of it is just incomplete. And the next version, whenever it arrives, is going to be stranger and more beautiful than anything we have drawn so far. Thanks for watching and I will see you in the next