We Found Galaxies Too Old for the Universe — And the Explanations Are Getting Wilder | Kip Thorne

Added: 2026-05-27

[00:00:00]There is a specific kind of vertigo that comes from looking at a photograph and realizing that the thing you are looking at should not be able to exist. Not in the aesthetic sense, not in the sense of finding something beautiful and wondering at its improbability.

[00:00:19]In the precise, mathematical, physically constrained sense that the object in the image violates the predictions of the best theory we have for how the universe works. that there was not enough time in the history of the cosmos uh for it to have uh grown to its current size. This is the vertigo that astronomers began experiencing in 2022 when the first science images from the James Webb Space Telescope started coming back to Earth. Not all the images, not even most of them, but a subset, a growing, persistent, stubbornly reproducible subset of images showing galaxies in the very early universe that appeared to be far too massive, far too developed, far too old for the time in which they existed.

[00:01:10]Galaxies containing billions or even tens of billions of times the mass of the sun in stars already fully formed and in some cases already dying in a universe that was supposedly too young to have had time to build them. The story of these galaxies, what they are, what they appear to mean, and what the scientific community has been doing to understand them is the subject of this documentary. It is a story that starts with the deepest images ever taken of the early universe and ends with a set of proposed explanations that point in directions so different from each other that some of them cannot all be right.

[00:01:49]It is a story about the limits of our best model of the universe about the difference between an anomaly that will eventually be absorbed and one that will require us to fundamentally revise something we thought we understood. And it starts, as so many of the most important stories in modern cosmology do, with the question of how we know how old a galaxy is and how we know how far away it is. The universe is expanding.

[00:02:15]This is the foundational observational fact of modern cosmology established definitively in the late 1920s by Edwin Hubble, building on earlier work by Vesto Slifer and others and confirmed with increasing precision by every subsequent generation of astronomical instruments. The expansion has a specific character that is crucial for understanding what the James Web Space Telescope is actually seeing when it images distant galaxies. It means that the farther away a galaxy is the faster it is receding from us and the faster it is receding from us, the more its light is redshifted, stretched toward longer wavelengths by the expansion of the space between us. This red shift is not like the Doppler shift of a moving ambulance. It is the stretching of the photon's wavelength by the expansion of space itself operating over the entire length of the photon's journey. A photon emitted by a galaxy when the universe was onetenth of its current size has been stretched to 10 times its original wavelength by the time it reaches us.

[00:03:26]Because the universe has expanded by a factor of 10 since the photon was emitted. We quantify this stretching with a number called the red shift Z, where a red shift of zero means the galaxy is effectively at the same point in cosmic time as us. And higher values of Z mean earlier times. A galaxy at redshift one is seen as it was when the universe was about half its current age.

[00:03:50]A galaxy at redshift 5 is seen as it was when the universe was about a billion years old.

[00:03:56]A galaxy at red shift 10 is seen as it was when the universe was only about 500 million years old, less than 4% of the current age. The Hubble Space Telescope over its three decades of operation pushed the frontier of deep galaxy observations to red shifts of about 10 to 11 with a few tenative candidates at higher red shifts. These observations showed a universe of young, mostly small, irregularlooking galaxies at high red shift, broadly consistent with theoretical models of early galaxy formation. The galaxies at the edge of Hubble's reach were mostly faint, compact, and not particularly massive.

[00:04:38]The building blocks of the larger, more settled galaxies that would emerge later. The James Webb Space Telescope was designed to push significantly beyond this frontier. JWST is the most powerful space telescope ever built with a primary mirror 6 and a half meters in diameter, nearly three times the diameter of Hubble's mirror, and instruments optimized for the near infrared wavelengths at which the light from very distant, highly redshifted galaxies arrives. The same physics that redshifts galaxy light to longer wavelengths as the universe expands means that the visible and ultraviolet light from early galaxies is shifted into the infrared by the time it reaches us. And JWST is exquisitly sensitive to exactly those wavelengths. It launched in December 2021 and began science operations in mid2022.

[00:05:36]Within months of its first science data release, the astronomical community was looking at something unexpected.

[00:05:44]The early images from JWST showed galaxies at red shifts beyond 10 in the universe's first billion years that were significantly brighter and more massive than the models had predicted. Not marginally more massive. not by factors of a few that might be explained by calibration uncertainties or modeling assumptions that could easily be tweaked. Some of the candidates appeared to be 10 to 100 times more massive than the models expected for galaxies at those red shifts and cosmic ages. One of the most discussed early findings was a set of six galaxy candidates analyzed by Evo Lab and his colleagues at Swinburn University of Technology published in Nature in early 2023.

[00:06:33]Lab A's team found that these six galaxies, if their phototric red shifts were correct, and if their mass estimates were right, contained between 10 billion and 100 billion solar masses of stars each in a universe that was only 500 to 700 million years old. The most massive of them appeared to contain more stellar mass than the Milky Way, despite the Milky Way having had 13 billion years to build its stars and these objects having had only a small fraction of that time. The phrase that went viral in the astronomical community and then in the popular science press was that these galaxies were impossible.

[00:07:17]Mike Boland Culchin, an astronomer at the University of Texas at Austin who had been working on the theoretical problem of early massive galaxies before JWST was even launched described the uh the situation in a paper published in April 2023 in Nature Astronomy. He wrote that if these phototric mass estimates were correct, the six galaxies collectively contained more stellar mass than all the stars that the standard cosmological model, the lambda CDM model predicted should exist in the entire observable universe at those red shifts.

[00:07:58]Not more than expected in those specific galaxies, more than predicted to exist anywhere. The statement was stark. It was also uh in the in the careful uh language of science conditional uh if the mass estimates were correct and that conditional is doing an enormous amount of work. Let me spend some time on it because it is at the heart of how the scientific community has been responding to the JWST early galaxy problem over the past three years and because understanding it is essential for evaluating how alarmed we should actually be about what the telescope is showing us. Measuring the mass of a distant galaxy is not a direct measurement. You cannot put a galaxy on a scale. What you do is measure the light coming from the galaxy, its spectral energy distribution, the amount of light at each wavelength, and then fit models to that light to infer the properties of the stellar populations generating it. The models encode what different mixtures of young and old stars with different metallicities and star formation histories look like in terms of their light output at different wavelengths. The best fit model gives you estimates of the stellar mass, the age of the stellar population, the current star formation rate and other properties. The problem is that these models are not perfect and the inferences they produce are not unique.

[00:09:35]Uh a galaxy's broadband phototric measurements the brightness in several different wavelength filters which is what JWST's imaging provides can be fitted by multiple different combinations of stellar population parameters. In particular, there is a known degeneracy between an old stellar population and a young stellar population with strong emission lines. A galaxy containing mostly old stars produces a specific spectral signature with most of its light at red wavelengths and little at blue wavelengths. But a galaxy containing mostly young hot stars can produce a similar red color if it also has strong emission lines from ionized gas that fall within specific filter band passes, boosting the the the apparent brightness in those filters and making the galaxy look redder and more massive than it actually is. The emission line contamination problem was flagged immediately by several groups as a significant potential source of bias in the initial JWST early galaxy mass estimates. If the anomalously massive galaxy candidates were actually bright because of emission lines from ionized gas around young hot stars rather than from a large mass of old stars, the mass estimates could be dramatically overestimated. A galaxy that looked like it contained 100 billion solar masses of old stars might actually contain only 10 billion solar masses of young stars if the broadband measurements were dominated by emission lines rather than stellar continuum light. The resolution of this ambiguity requires spectroscopy.

[00:11:21]Not just measuring the total brightness in broad wavelength bins, but actually dispersing the light into a spectrum and measuring individual features at specific wavelengths. Emission lines appear at known wavelengths determined by atomic physics. And if they are present and contributing significantly to the broadband measurements, they will be clearly visible in a spectrum. The mass estimate from spectroscopy is far more reliable than from photometry alone because the spectroscopic information breaks the degeneracies that photometry cannot resolve. JWST has spectroscopic capabilities. Its nearspec instrument is one of the most powerful near infrared spectrographs ever deployed in space and the follow-up spectroscopic observations of the JWST early galaxy. candidates began almost as soon as the phototric findings were published. The results have been in the language appropriate to this situation mixed. And by mixed I mean genuinely informative but pointing in multiple directions at once which is exactly the situation that makes a scientific problem interesting and exactly the situation that makes writing a clean and definitive account of it difficult. Here is what the spectroscopy has shown. Some of the phototric galaxy candidates turned out to be at much lower red shifts than the photometry suggested they were not in the early universe at all. The phototric redshift technique, which infers red shift from the color of a galaxy's light, can produce wrong answers when specific combinations of dust reening and spectral features mimic the signatures of high red shift galaxies. When spectra were obtained for some of the original lab at all candidates, um, a fraction of them turned out to be lower redshift interlopers.

[00:13:21]Not most of them, the fraction of catastrophic phototric redshift failures in this specific data set turns out to be smaller than pessimism feared, but enough to reduce the effective sample size and somewhat mitigate the tension with the standard model. For the candidates that spectroscopic observations confirmed were indeed at high red shift, the spectroscopic mass estimates were in most cases lower than the phototric estimates as the emission line contamination hypothesis predicted.

[00:13:52]The confirmed high redshift galaxies are massive, significantly more massive than naive prejwst models would have predicted, but not as extreme as the initial phototric estimates suggested.

[00:14:04]The revised stellar masses after spectroscopic confirmation and emission line correction typically come out a factor of 3 to five lower than the phototric estimates which is meaningful.

[00:14:17]It reduces the claimed tension with the standard model by a corresponding factor. This is essentially what happened with NASA's announcement in August 2024 that the problem of seemingly too early massive galaxies was largely a consequence of confusion with active galactic nuclei with rapidly accreting super massive black holes rather than pure stellar mass. The NASA analysis based on spectroscopic and multi-wavelength follow-up of several of the most alarming initial candidates found that a significant fraction of the anomalously bright objects were not dominated by starlight but by the light from accretion discs around feeding super massive black holes which can be orders of magnitude more luminous per unit mass than stellar populations and which produce spectral signatures that can be confused with old stellar populations in the limited spectral information of phototric measurements.

[00:15:16]When the black hole contribution is removed, the underlying stellar masses of these objects become much more consistent with what the standard model predicts. The headline from this line of work is the problem is less severe than the initial phototric results suggested.

[00:15:36]not gone, not solved, less severe. But here is why the story is not over and why the early galaxy problem continues to be one of the most actively discussed topics in observational cosmology as of 2025.

[00:15:52]Because even after correcting for phototric biases, emission line contamination and active galactic nucleus contamination, there remains a population of spectroscopically confirmed mass reliable high red shift galaxies that are more massive than the standard model predicts. Not 10 times more massive as the initial phototric candidates appeared, but factors of a few, two, three, five more massive than the models expect at those red shifts.

[00:16:26]And that factor of a few discrepancy when you understand what the standard model is predicting and why is genuinely significant. To understand why, I need to explain what the standard model of cosmology actually predicts about how galaxies form in the early universe and why the predictions it makes are specific rather than vague. The standard model of cosmology, the lambda CDM model, where lambda refers to the cosmological constant that drives the accelerating expansion of the universe and CDM refers to cold dark matter is not merely a qualitative description of how the universe evolved. It is a quantitative model with specific parameters that have been measured with extraordinary precision primarily from the cosmic microwave background radiation. the afterglow of the big bang which plank and wap and other instruments have mapped with exquisite detail. These parameters specify the composition of the universe at early times, the spectrum of density fluctuations that seeded the large scale structure, the rate of expansion at every epic, and the total amount of matter and energy in various forms.

[00:17:40]Given these parameters, it is possible to run numerical simulations of cosmic evolution to start with the conditions of the early universe and calculate using the laws of physics as we understand them how matter should clump together under gravity, how dark matter halos should form and merge, and how gas inside those halos should cool and collapse to form stars and galaxies. The most sophisticated versions of these simulations, Illustrous TNG, Fire, Eagle, and others are extraordinary computational achievements. Running on some of the world's largest supercomputers and tracking hundreds of millions of virtual particles through billions of years of simulated cosmic time. These simulations make specific testable predictions for the masses of the most massive galaxies that should exist at every epoch in cosmic history.

[00:18:36]And the predictions arise from a specific physical logic. You can only form a galaxy if the dark matter halo that hosts it has had time to grow to a sufficient mass for its gravity to attract and retain gas and for that gas to have had time to cool and collapse into stars. In the early universe, when the cosmos was young and dark matter halos were small, the most massive halos that could have formed by any given time were constrained by the hierarchical assembly process of structure formation.

[00:19:12]Large halos form from the merging of smaller halos and there is only so fast that small halos can merge into large ones given the fundamental constraint that the information about where the dark matter is cannot travel faster than gravity which propagates at the speed of light. The result is a specific prediction. At red shift 10, the universe at 500 million years old, the most massive galaxy that should plausibly exist has a stellar mass of perhaps a few hundred million solar masses with the most extreme outliers in exceptional circumstances, reaching perhaps a billion solar masses. 10 billion solar masses is what you might expect in the Milky Way after 12 billion years of growth. Finding anything close to 10 billion solar masses at red shift 10 is under the standard model extremely unlikely. Finding it routinely in multiple objects is not just improbable.

[00:20:13]It is if the measurements are right impossible in the strict mathematical sense that no combination of standard physics can produce the required structures in the available time. So the key question becomes which is wrong the measurements or the model. The history of science offers many examples of both possibilities.

[00:20:38]There have been many cases where anomalous observations turned out to be measurement errors, systematic biases or misidentified objects. The initial fast pulsar candidates the faster than light nutrino result from opera. the claimed detection of gravitational waves by bicep 2 before the dust contamination was understood. The track record of anomalous results ultimately surviving rigorous scrutiny and pointing to genuine failures of the standard model is considerably shorter, but it exists.

[00:21:14]The discovery of the accelerating expansion of the universe which required adding the cosmological constant back into the standard model is the most dramatic recent example for the JWST early galaxy problem. The community has been exploring both possibilities simultaneously and the directions they point in are genuinely diverse. Let me walk through the major proposed resolutions because each one reveals something interesting about what might be happening in the early universe. The first category of resolution is observational. The mass estimates are wrong and when corrected the tension disappears. This is the position that the NASA analysis of active galactic nucleus contamination supports most strongly and it is the most conservative and least paradigmth threatening conclusion. Under this view, the initial JWST phototric candidates were contaminated by emission lines, AGN light or other astrophysical confusion.

[00:22:19]and the spectroscopic followup when complete will reduce the apparent masses to values consistent with the standard model. The argument for this position is that the history of high red shift galaxy observations is replete with exactly these kinds of corrections.

[00:22:39]Initial estimates are systematically high because the most massive looking objects tend to be the most confused ones. and uh follow-up observations consistently bring the masses down. The argument against this position as a complete resolution is that even after all currently known systematic corrections are applied, there remains a population of spectroscopically confirmed galaxies that are too bright and too massive for their red shift. The corrections reduce the tension but do not eliminate it. And the specific galaxy that has become perhaps the most discussed individual case, JWST 7329, published in Nature in February 2024 by a team led by Claudia Logos at the International Center for Radioastronomy Research seems to be particularly difficult to explain away by measurement error. JWST 7329 is a massive old passive galaxy. old in the sense that it appears to have stopped forming stars observed at a red shift corresponding to about three billion years after the big bang. Its stellar population analysis suggests it formed most of its stars in the first billion years of the universe around red shift 10 or higher. What makes JWST 7329 especially puzzling is not just its mass. It is that its stellar mass appears to exceed the total mass of the dark matter halo that should have been hosting it at the time of its formation.

[00:24:14]If standard dark matter halo models are correct, the galaxy seems to have formed more stars than the dark matter structure available to see it could have supported. This is not merely a tension with models of star formation efficiency. It is a tension with the underlying dark matter structure formation model itself with the predictions of how dark matter halos grow in the early universe. The second category of resolution is astrophysical.

[00:24:42]The standard model of galaxy formation is wrong in specific ways but the underlying cosmological model is correct. Under this view, the dark matter halo mass function, the distribution of how many halos of different masses exist at each epic is correct. But the relationship between halo mass and galaxy mass is different in the early universe from what the simulations predict. Specifically, star formation in the early universe may be far more efficient than the simulations assume. In the standard simulations, a galaxy typically converts somewhere between two and 20% of the gas in its host dark matter halo into stars over its lifetime. The rest is prevented from forming stars by feedback processes.

[00:25:29]Supernova explosions that blast gas out of galaxies. Radiation from young hot stars that photoionizes and heats the surrounding gas. jets from active galactic nuclei that heat and expel gas from galaxy halos. These feedback processes are included in the simulations, but they are modeled with limited precision because they operate on spatial scales far too small to be directly resolved in cosmological scale simulations. They are instead represented by simplified prescriptions, recipes that capture their average effects but may not accurately reproduce their detailed behavior in the specific conditions of the early universe. In the early universe, the conditions were different from today in ways that could plausibly affect star formation efficiency. The gas was denser because the universe was younger and more compact. The metallicity was lower.

[00:26:22]There were fewer heavy elements because stellar nucleiosynthesis had not had time to enrich the gas significantly and this affects how efficiently the gas cools and collapses to form stars. The first generation of stars known as population 3 stars were likely much more massive than typical stars today because the absence of metals changes the cooling processes that regulate the minimum mass of stars. The feedback from these massive first stars may have been different from the feedback that the simulations calibrated on lower red shift galaxies. If the star formation efficiency in the first billion years of the universe was significantly higher than the simulations assume, if the early universe was in some specific conditions able to convert a larger fraction of its gas into stars than the standard models predict, then the observed high red shift galaxy masses could be consistent with the standard dark matter halo distribution. The halos would be the right size. the galaxies would just be more efficient at using the material in those halos. Several groups have explored this possibility and found that it can work quantitatively if the star formation efficiency is pushed to values near 100% in the most massive early halos.

[00:27:44]100% means all the gas in the halo converts to stars with no gas lost to feedback. This is an extreme value. It requires the feedback processes to be essentially ineffective in these objects which would need specific physical explanation but is not physically impossible. A 2023 paper by Decel and colleagues proposed a specific mechanism. In the early universe, the rapid inflow of cold gas into dense halos could create conditions where star formation proceeds so quickly that the stars form before supernova feedback has time to disrupt the gas reservoir. The gas comes in, forms stars rapidly, and the feedback from those stars arrives too late to prevent the next wave of gas infall and star formation. The result is a runaway star formation episode that produces a galaxy far more massive than the models would predict from a halo of that size at that time. This is a testable prediction. Galaxies formed through this cold stream dominated rapid assembly mechanism should have specific chemical and morphological properties.

[00:28:55]high metalicities from the rapid cycling of material through stars, compact morphologies from the dense gas infall, and possibly rotation signatures from the angular momentum of the infalling streams. JWST has the capability to measure these properties for the brightest confirmed high red shift galaxies and the spectroscopic programs to do so are currently underway. The results will be available over the next few years and will test the cold stream hypothesis directly. The third category of resolution is cosmological. The standard model of cosmology itself needs revision. This is the most radical and most scientifically consequential possible resolution and it is the one that has generated the most theoretical creativity and the most controversy. The standard lambda CDM model has been extraordinarily successful. It explains the cosmic microwave background temperature fluctuations to exquisite precision. It correctly predicts the abundances of light elements produced in big bang nucleiosynthesis.

[00:30:03]Uh it accounts for the large scale structure of the universe, the distribution of uh galaxies and galaxy clusters on scales of of hundreds of millions of light years. It correctly describes the accelerating expansion of the universe. Its predictions have been tested against observations across a range of scales and epochs and have been found to be extraordinarily accurate.

[00:30:27]And yet it has problems. The Hubble tension, the discrepancy between the expansion rate of the universe measured from the cosmic microwave background and the expansion rate measured from nearby distance indicators has been growing in statistical significance for years and remains unresolved. The S8 tension, a discrepancy between the predicted and observed amplitude of matter clustering on large scales, has also been accumulating. These are real quantitative discrepancies that could indicate that something in the standard model is slightly wrong in ways that have not yet been identified. The JWST early galaxy problem could be another manifestation of the same underlying incorrectness or it could have a completely different explanation. The theoretical proposals for cosmological modifications that could resolve the early galaxy problem are numerous and varied and I want to describe several of them in some detail because they represent genuinely different visions of what the universe might fundamentally be. One class of proposals involves primordial black holes. In the standard model, the first seeds for super massive black holes were likely stellar mass black holes left over from the deaths of the first generation of stars, which then grew by accreting gas over billions of years. This standard black hole formation pathway has difficulty producing the super massive black holes with masses of billions of solar masses that JWST has found in galaxies at red shifts above seven because there is not enough time for a stellar mass black hole to grow to those masses through standard accretion. But if some black holes formed in the very early universe from density fluctuations during the inflationary phase before the big bang rather than from stellar evolution, they could have started much more massive.

[00:32:28]These primordial black holes could have then grown faster and also served as seeds for galaxy formation, pulling in gas and dark matter at earlier times than standard stellar black holes could have. The presence of primordial black holes would change the galaxy mass function at high red shift in ways that could match what JWST is seeing. The primordial black hole hypothesis has the advantage of also addressing the super massive black hole problem separately from the galaxy mass problem and it connects to a broader set of observational tests. The gravitational wave signal from primordial black hole mergers would be detectable by the laser interferometer space antenna which is being developed by the European Space Agency and is planned for launch in the 2030s. The specific mass distribution and merger rate of primordial black holes would leave signatures in the LISA data that are distinguishable from those of black holes formed through stellar evolution. The second class of proposals involves modifications to the nature of dark matter. In the standard model, dark matter is cold. It consists of particles that were moving slowly relative to the speed of light when they decoupled from the rest of the universe in the early hot plasma. And this slowness means the dark matter could begin forming dense clumps on small scales very early. If dark matter were instead warm, consisting of particles with somewhat higher velocities, at decoupling, the formation of the smallest dark matter structures would be suppressed because the free streaming of the particles would wash out smallcale density fluctuations.

[00:34:10]Warm dark matter models have long been studied as an alternative to cold dark matter, primarily motivated by the apparent deficit of small galaxies in the standard model relative to observations. The JWST early galaxy problem adds a new constraint. The dark matter model needs to produce not fewer small structures than cold dark matter but more massive large structures in the early universe. Uh this is actually the opposite tendency from warm dark matter which makes the early galaxy problem harder to solve through warm dark matter modifications. Some researchers have therefore turned in the opposite direction to models with enhanced small-cale power where the dark matter produces more dense structures at small scales in the early universe than cold dark matter predicts. These modifications can increase the abundance of massive halos at high red shift providing more environments for efficient star formation and potentially resolving the tension.

[00:35:12]uh a third class of proposals is more radical. Perhaps the age of the universe is different from what lambda CDM implies. The standard model places the age of the universe at approximately 13.8 billion years based on the Hubble constant, the matter density, and the cosmological constant. If any of these parameters were different, the age would be different. A universe that is older, perhaps 26 or 27 billion years old, as proposed in some modified cosmological models, would have had more time to form massive galaxies by any given red shift, potentially eliminating the problem entirely. One such proposal is the covering coupling constants model developed by Rejendra Gupta at the University of Ottawa which modifies the standard Freriedman equations by allowing the fundamental constants of nature to vary over cosmic time and which produces a universe age of approximately 26.7 billion years. In this model, there is ample time at red shift 10, roughly 5.8 billion years into the universe's history. Rather than the standard 500 million years for massive galaxies to form through conventional processes, the model fits the supernovi type IIA distance data reasonably well and is explicitly designed to resolve the JWST early galaxy problem. Critics of the proposal note that varying fundamental constants must be consistent with a wide range of other observational constraints and theoretical considerations and that the specific variation proposed by Gupta is not well motivated from particle physics. But the paper has been published and cited and represents one direction in which the theoretical response to JWST has gone. uh a more mainstream theoretical response, one that does not require either primordial black holes or varying constants, focuses on early dark energy. The Hubble tension, the discrepancy between high red shift and low red shift measurements of the expansion rate has motivated a class of models in which an additional component of dark energy was present in the early universe. dark energy that temporarily increased the expansion rate around the epoch of matter radiation equality and then decayed away, leaving a universe that has since evolved approximately as lambda CDM predicts.

[00:37:43]Early dark energy models were originally proposed to resolve the Hubble tension.

[00:37:48]And they work by modifying the sound horizon in the cosmic microwave background in a way that changes the inferred Hubble constant from CMBB measurements to better match low red shift determinations.

[00:38:03]The interesting recent development is that early dark energy also helps with the galaxy mass problem. If the expansion rate was faster in the early universe than standard lambda CDM predicts, the largecale structure would grow differently. And in some models, the density contrasts the degree to which overdense regions were overdense relative to the background would be enhanced in ways that produce more massive dark matter halos earlier. A 2024 analysis by Xiao and colleagues showed that a specific early dark energy model that fits both the Hubble tension and the galaxy mass discrepancy simultaneously could be constructed with the two effects arising from the same underlying modification to the expansion history. This is an appealing unification solving two apparently unrelated problems with a single theoretical innovation. Though it remains to be seen whether the specific model parameters required survive the full range of cosmological constraints.

[00:39:08]What all of these proposed resolutions have in common is that they are reaching beyond the standard picture in some way.

[00:39:17]Either adjusting the astrophysics within a fixed cosmological framework or modifying the cosmological framework itself. None of them is obviously correct. None of them has been definitively ruled out. And the observations that would distinguish between them detailed spectroscopic characterization of more high redshift galaxies, measurements of the galaxy's stellar mass function at red shifts beyond 12, better constraints on the Hubble tension and the S8 tension, gravitational wave observations that could probe primordial black holes are all underway or planned for the next decade. This is the specific scientific situation as of mid2026.

[00:40:00]A genuine tension between observations and the standard model with a range of proposed explanations whose relative plausibility is being refined by an ongoing flood of new data from JWST and other facilities. The tension has been significantly reduced from the initial most alarming phototric estimates, but it has not been eliminated. And the most carefully studied individual objects, particularly the confirmed spectroscopic cases of early massive passive galaxies like JWST 7329, remain genuinely difficult to explain within the standard uh framework without some modification.

[00:40:40]Let me now describe what JWST is actually seeing in more detail. Because the images and spectra deserve closer attention than the mass estimates alone provide. The early universe revealed by JWST is not simply a place full of unexpectedly massive galaxies. It is a place that is surprising in multiple somewhat independent ways. Each adding to a picture of the first billion years of cosmic history that is richer and stranger than the preJWST picture in ways that go beyond the galaxy mass question. One of the most striking findings is the presence of galaxies that have already stopped forming stars.

[00:41:19]What astronomers call quiescent galaxies at red shifts above four and five in the first two billion years of cosmic history. Galaxy quenching the process by which a star forming galaxy transitions to a passive non-star forming state is one of the most important and least understood processes in galaxy evolution. On Earth's cosmic time, the majority of quenched galaxies are found in dense clusters and appear to have been quenched by environmental processes related to their cluster membership or by feedback from super massive black holes that heat and expel the gas reservoirs needed for star formation.

[00:42:04]The JWST finding of quiescent galaxies at red shift above four means that quenching was already operating in the first two billion years far earlier than expected. Particularly striking is a galaxy called GS 9209 confirmed at a red shift of about 4.6 six, roughly 1 billion years after the Big Bang, which had already formed the equivalent of the Milky Way's stellar mass and then stopped forming stars entirely. Its star formation having been quenched in a relatively short period.

[00:42:40]The mechanism that quenched it in such a brief cosmic time is uncertain, but the leading candidate is feedback from an active galactic nucleus, a super massive black hole growing rapidly and driving winds that expelled the galaxy's gas.

[00:43:01]The black hole that would have done this must itself have grown very rapidly to the mass needed to drive effective quenching winds, adding to the evidence that something in the standard picture of black hole growth needs revision. The finding of very early quiescent galaxies is potentially more constraining than the finding of very early massive star forming galaxies for the following reason. A massive star forming galaxy at high red shift at least requires only that its stars formed. The formation of stars while requiring efficient gas cooling and collapse does not absolutely require any specific earlier phase. But a quiescent galaxy at high red shift requires not just that a large mass of stars formed but that after forming the galaxy's star formation was then shut off by some process. That process takes time and specific physical conditions to operate. Finding quiescent galaxies in the first billion years means that the full cycle formation, peak star formation, quenching completed in a small fraction of the time that this cycle takes in the local universe where it typically requires several billion years. The physics of quenching must work faster in the early universe than the standard models predict. Then there is the finding from uh 2026 of a galaxy that formed uh less than two billion years after the big bang with a mass several times that of the Milky Way but shows no rotation whatsoever. The non-rotating galaxy, published in late May 2026 by a team led by Ben Forest at UC Davis, is puzzling because essentially all large galaxies in the local universe that have no rotation are elliptical galaxies, the products of major mergers and significant dynamical evolution that took many billions of years to transform.

[00:45:04]Rotating discs into pressured supported non-rotating systems. Finding a non-rotating massive galaxy at cosmic ages of only one to two billion years suggests either that the dynamical evolution proceeded extraordinarily quickly or that this galaxy never had a rotating disc phase to begin with assembling directly into a pressure supported spheroid through processes that the standard model does not predict at this cosmic age. Each of these JWST findings is uh individually a challenge to some aspect of the standard picture.

[00:45:46]Together they paint a portrait of the early universe that is consistently though not catastrophically at odds with what the models predicted. The early universe appears to have been more efficient at forming massive galaxies, more effective at quenching them, more capable of producing dynamically mature systems, and more diverse in its galaxy population than our preJWST simulation suggested. The scientific community's response to this situation has been on balance the appropriate one. There has been genuine engagement with the data, serious analysis of systematic effects, a range of proposed explanations spanning from conservative to radical, and an ongoing program of follow-up observations designed to discriminate between the possibilities. No credible scientist has said that the standard model is definitively wrong. No credible scientist has said that the JWST findings are definitely the result of measurement error. The honest position, the one most reflective of the actual state of knowledge is that the tension is real, its magnitude is uncertain, and its ultimate cause is not yet established. What makes this moment particularly interesting from a historical perspective is the comparison to previous episodes in cosmology when anomalous observations forced theoretical revision. The discovery of the accelerating expansion of the universe in 1998 from type IA supernova distance measurements was initially received with skepticism because it required the reintroduction of the cosmological constant that Einstein had famously added and then abandoned. But the data was solid, the systematic effects were manageable, and the conclusion was confirmed by multiple independent lines of evidence. The cosmological constant reinterpreted as dark energy is now a standard component of the lambda CDM model. The JWST early galaxy problem is not at the same stage of certainty as the accelerating expansion was in 2003. After independent confirmation from the cosmic microwave background and largecale structure data, the JWST findings are genuinely uncertain. The systematic effects are larger. The sample sizes are smaller.

[00:48:14]The alternative explanations are more numerous and more viable. But the direction of the findings is consistent.

[00:48:20]The early universe has more massive, more developed, more dynamically mature galaxies than the standard model predicts. And the persistence of this pattern across multiple independent analyses, multiple observing programs, multiple galaxy samples studied by different groups suggests that at least some component of the tension will survive the full accounting of systematic effects. Whether that surviving component will require a modification to the cosmological model or can be absorbed within a significantly revised picture of early galaxy formation within the standard cosmological framework is the central question and it is a question that JWST itself is positioned to help answer as the observing programs designed to characterize early galaxies spectroscopically and to push the phototric redshift frontier. here to even higher red shifts continue to accumulate data. The telescope has already pushed the confirmed redshift frontier to unprecedented depths. In April 2023, the galaxy J A DGSC130, spectroscopically confirmed at a red shift of 13.2, 2 became the most distant spectroscopically confirmed galaxy at the time of its announcement observed as it was at 320 million years after the Big Bang when the universe was roughly 2% of its current age. This record has since been extended further. The identification and spectroscopic confirmation of galaxies at red shifts above 13 is now routine for JWST, representing a capability that was completely out of reach for Hubble. And that is revealing a universe in its infancy in the first few hundred million years after the Big Bang that no previous telescope could study in detail. What JWST is showing us from this early epoch is the universe in the process of becoming what it is. The first stars, population three stars formed from pristine primordial hydrogen and helium without any heavy elements are not directly detected yet. But their descendants are being observed in the enriched gas of the first few hundred million years. The first galaxies are not simple, smooth, featureless clumps of gas condensing gradually under gravity. They uh are already showing uh signs of complexity, multiple stellar populations, chemical enrichment, structural development, and in some cases the signatures of nuclear activity that indicate growing uh super massive black holes in ways that suggest the formation of the first cosmic structures was a more rapid and more violent process than the gradual hierarchical assembly of lambda CDMs smoothly. ly evolving density fluctuations implies.

[00:51:26]Whether this is because the density fluctuations were not as smooth as the standard model assumes, perhaps with enhanced smallcale power from modified dark matter physics or primordial features in the inflation spectrum, or because the astrophysics of star and galaxy formation in the early universe is more extreme than the simulations reproduce, or whether it requires more fundamental revision to the cosmological model. is the question that will occupy a generation of cosmologists.

[00:51:57]The history of cosmology is a history of models being extended, revised, and occasionally overturned by observations that exceeded what the previous model could accommodate. Newton's mechanics were extended by special relativity, which was extended by general relativity. The static universe model was replaced by the big bang model. The pure matter dominated big bang model was extended to include inflation which solved the horizon and flatness problems. The matter dominated expanding universe was supplemented with dark energy when the accelerating expansion was discovered. Each revision preserved the successful predictions of the previous model while incorporating the new observations into an updated framework. The Lambda CDM model with its extraordinary track record of successful predictions is not going to be abandoned in favor of any of the currently proposed alternatives on the basis of the JWST early Galaxy findings alone. Even if the tension proves to be real and significant, what is more likely based on historical analogy is that the model will be extended supplemented with additional physics that modifies its predictions for the early universe while preserving its predictions for the better tested epochs and scales in a way that incorporates whatever is genuinely new about the JWST findings without discarding what has been established over decades of successful cosmological observation.

[00:53:35]What that extension looks like, whether it involves primordial black holes, modified dark matter, early dark energy, enhanced star formation efficiency, or some combination of these and other possibilities is what the next decade of observations and theoretical work will determine. In the meantime, we have something genuinely extraordinary. The most powerful telescope in human history is looking at the universe in its infancy. And what it is showing us is that the beginning was stranger and more complex and more productive of structure than we expected. The universe formed its first massive galaxies faster than our models predicted. It quenched some of them faster. It produced rotating discs and non-rotating spheroids and active galactic nuclei and complex chemical enrichment patterns in the first billion years in what our models had described as the simple early precomplex universe of the first generation of structures. The universe was busy from the beginning, more busy and more creative and more capable of rapid complexity than the models said it should be. That is not nothing. That is in fact one of the most interesting things that has happened in observational cosmology in decades.

[00:55:01]and um uh the James Webb Space Telescope floating a million miles from Earth at the stable gravitational balance point called L2, cooling its instruments to within a few degrees of absolute zero, capturing photons that have been traveling for 13 billion years, is only in the early years of its planned mission. The data it has collected so far represents a small fraction of what it will eventually produce. The galaxies are too old for the universe we thought we lived in or the universe we thought we lived in was too simple for the galaxies it contained. We are in the process right now of finding out which the answer will tell us something fundamental about the beginning of everything and we will have it in all its complexity and all its clarity within the next decade. The universe does not hide its history forever. James Webb is finding the pages we missed. Let me go deeper now into the specific science of how astronomers are attempting to weigh the competing explanations because the methodology is as interesting as the conclusions and because understanding how you distinguish between the astrophysics is wrong and the cosmology is wrong is one of the more intellectually demanding tasks in modern observational science.

[00:56:28]The fundamental challenge is that almost all of the explanations that have been proposed are not mutually exclusive. Uh a universe uh with primordial black holes also has galaxies in it. A universe with enhanced star formation efficiency in the early epochs also has normal dark matter halos. A universe with early dark energy also has the same basic structure formation physics. The different explanations differ in their predictions for the magnitude of the excess and for the specific properties stellar ages, chemical compositions, morphologies, black hole masses of the galaxies that are more massive than expected. Discriminating between them requires not just knowing that the galaxies exist, but knowing in detail what they are made of and how they got to be that way. The most powerful tool JWST has for this discrimination is spectroscopy of individual stars and stellar populations at high red shift. A capability that was essentially impossible before the telescope launched and is now producing results that are transforming the field at a pace that is difficult to keep up with even for the scientists directly involved. When you take a spectrum of a high redshift galaxy with JWST's nearspec instrument, you are measuring the light at hundreds or thousands of individual wavelengths across a range that spans from about 0.6 to 5.3 micrometers.

[00:58:09]In that spectrum, you can identify specific spectral features, emission lines from ionized gas produced by star formation or AGN activity, absorption lines from specific atomic transitions in stellar atmospheres, the Balmer brake and the 400 angstrom break that are produced by specific stellar population age and metalicity combinations.

[00:58:35]and uh the lyman break that marks the wavelength below which the interstellar medium absorbs essentially all photons.

[00:58:42]From these features combined with the overall spectral shape, you can reconstruct the star formation history of the galaxy, the rate at which it was forming stars at each point in its past.

[00:58:55]You can measure the metallicity, the abundance of heavy elements, which encodes information about how many generations of stars the material has been through. You can measure the dust content, the ionization state of the gas, the presence or absence of AGN signatures.

[00:59:12]And uh you can do all of this at wavelengths that correspond to rest frame ultraviolet and optical light, which is where the most informationrich spectral features from stars and ionized gas live. Even though the light has been redshifted into the infrared by the time it reaches JWST.

[00:59:32]The picture that is emerging from the first few years of JWST spectroscopy of high red shift galaxies is a portrait of the early universe that is consistently surprising in the same direction. Things happen faster than expected. Star formation rates in some early galaxies are extraordinarily high. hundreds to thousands of solar masses of new stars per year compared to a few solar masses per year in the current Milky Way.

[00:59:58]Chemical enrichment happened quickly.

[01:00:00]Some galaxies at red shift five and six already have metalicities, a substantial fraction of the solar value, implying that many generations of stars had already lived and died. And the transition from active star formation to quiescence appears to have happened on time scales of a few hundred million years in some galaxies compared to the several billion years typical in the local universe. One of the most striking specific findings from JWST spectroscopy is the detection of strong nebular emission lines from very high red shift galaxies lines from oxygen, nitrogen, carbon, and hydrogen at abundances and ionization states that indicate the presence of young very hot stars. The presence of these emission lines was expected at some level since the first galaxies should be forming stars actively. What was not expected is how extreme some of these systems are. A handful of galaxies at red shift above 8 show emission line ratios that indicate an ionization field far harder with more high energy photons than typical star forming regions in the local universe.

[01:01:15]These are the conditions you would expect if the early galaxies contained a large population of extremely massive stars. Stars 30 to 100 times the mass of the sun or more. Or if the stars had anomalously low metalicities that made them hotter and more ionizing than typical stars today. This finding connects to one of the deepest questions in early universe science. What were the first stars like? The theoretical prediction based on the physics of gas cooling in a metal-free environment is that the first generation of stars, population three stars, formed from the pristine primordial gas, were extremely massive by the standards of later star formation with typical masses perhaps in the range of dozens to hundreds of solar masses. Such massive stars would be extremely luminous, extremely hot and would produce copious ionizing radiation. Their lifetimes would be short a few million years and their deaths would be dramatic core collapse supernovi hypernovi or in the most massive cases pair instability supernovi that completely destroy the star and disperse its enriched material across enormous volumes. JWST has not yet directly detected population 3 stars. They are likely too faint and too numerous, but small to be resolved as individual stars at any accessible red shift. But the evidence for their influence is present in the extreme ionization signatures of some high redshift galaxies and in an extraordinary finding published in October 2024 from a galaxy called Glass Z12. In this galaxy, JWST detected what the research team described as a potential missing link to the first stars, an unusual emission line signature that the team interpreted as evidence for a gas cloud being irradiated by a very hard radiation field. consistent with population three star characteristics. This is not yet confirmed as a definitive detection of pop three stars, but it represents the closest anyone has come and it points toward what the early universe genuinely looked like in the environments where the first stars were forming. The race to detect the first population three stars or at minimum to characterize the conditions in which they lived is one of the most exciting frontiers in current JWST science. If they are detected, it will be a fundamental milestone, the direct observation of the first objects to form from the primordial gas, the ancestors of all the heavy elements in the universe today. Everything heavier than lithium, the carbon in your DNA, the oxygen you breathe, the iron in your blood was first produced in the nuclear furnaces of stars. And the first furnaces were the population three stars that lived and died in the first 100 million years of cosmic history. To see them even indirectly through their effects on surrounding gas would be to look at the origin of the chemistry that makes life possible. There is another dimension of the JWST early galaxy findings that deserves careful attention. The discovery of extremely compact massive galaxies at high red shift and uh what their sizes and structures tell us about how galaxies grow in the local universe. massive elliptical galaxies, the large pressure supported non-star forming systems that represent the end points of many galaxies evolutionary paths are physically large objects. The most massive ellipticals in today's universe have effective radi of 10 to 20 kiloparex or more. Meaning they extend for tens of thousands of light years in every direction. Their sizes are consistent with the theoretical prediction that galaxies grow through mergers. When two galaxies merge, the resulting galaxy is typically larger than either progenitor and the most massive local galaxies are the products of many such mergers over billions of years. At high red shift, this picture should manifest as smaller galaxies because the high red shift galaxies are seen before they have had time to grow through mergers. And indeed, the Hubble Space Telescope observations of high redshift galaxies did show a trend toward smaller sizes at higher red shifts, broadly consistent with the merger-driven growth picture.

[01:06:09]JWST has extended this measurement to much higher red shifts. And what it is finding is that at the highest red shifts beyond Z of eight or nine, many massive galaxies are extraordinarily compact. The sizes measured for some early massive galaxies are on the order of hundreds of parexs less than a kiloparc across compared to the tens of kiloparex of their local universe equivalents. These are galaxies that if they are truly the ancestors of today's massive ellipticals will need to grow by factors of 10 to 50 in physical size over the next 12 billion years through mergers. The implied merger history is dramatic and specific and it should leave observable signatures in the largecale structure of the universe and in the statistical properties of galaxy mergers at intermediate red shifts. But there is another possible interpretation of the compact high red shift galaxies that has been gaining traction. Some of them might not be the ancestors of the largest local ellipticals. They might be systems that remain compact that never undergo the extensive merging required to grow to large sizes and that persist as compact relics in the local universe.

[01:07:22]Local universe surveys have indeed found a population of massive compact old galaxies called compact massive relic galaxies or red nuggets that appear to have survived the cosmic epochs without significant growth through merging.

[01:07:39]These systems are rare in the local universe, but their existence as a population is consistent with the idea that not all high red shift compact massive galaxies undergo the extensive merger history that would make them large. The balance between the two pathways, compact galaxies that grow into large ellipticals through mergers versus compact galaxies that remain compact as relics is a question that JWST is uniquely positioned to address.

[01:08:09]because it can characterize both the sizes and the stellar population properties of the high red shift compact galaxies with enough precision to trace their evolutionary histories. The spectroscopic ages of the stellar populations in these compact objects combined with measurements of their chemical abundances and internal kinematics will eventually paint a complete picture of how the earliest massive galaxies formed and evolved. I want to spend a few minutes now on one of the most philosophically interesting aspects of the JWST early Galaxy problem because I think it reveals something important about the nature of scientific knowledge and the specific kind of challenge that the JWST findings represent. The Lambda CDM model was not constructed in isolation from observational constraints. It was built piece by piece over decades in response to the accumulation of cosmological observations from multiple independent probes. The value of the Hubble constant was determined from sephiid distances and supernova observations. The matter density was determined from largecale structure surveys. The dark energy density was determined from supernova distances and the CMB. The primordial density fluctuation spectrum was measured from the CMBB temperature anosotropies.

[01:09:40]The model was not proposed theoretically and then confirmed observationally. It emerged from the data with its parameters tuned to match the observations at each epic. This means that when JWST finds something that the lambda CDM model does not predict, the question is not simply is the model wrong. The question is which aspect of the model is wrong and how does changing that aspect affect all the other observations that the model correctly describes. A modification to the model that resolves the JWST early galaxy problem must not break the model's successful predictions for the cosmic microwave background for the barrier and acoustic oscillation measurements for the large scale matter distribution for the type IA supernova distances. The space of allowed modifications is constrained not by a single observational inconsistency but by the requirement of consistency with the full ensemble of cosmological observations.

[01:10:41]This is why cosmological modifications proposed to resolve the JWST tension are so carefully constrained and so technically elaborate. Early dark energy models must be calibrated to affect the expansion history at the right epoch without distorting the CMB acoustic peaks. Primordial black hole scenarios must not overproduce gravitational wave backgrounds detectable by current or planned detectors.

[01:11:08]Modified dark matter scenarios must reproduce the correct large-scale structure without violating constraints from small-scale structure observations.

[01:11:18]Each proposed resolution exists within a web of constraints that limits how far it can deviate from the standard picture. The fact that multiple independent proposed resolutions astrophysical and cosmological all seem capable of reducing the JWST tension without obviously violating other constraints is itself informative. It means the tension is real but its magnitude is insufficient to uniquely identify the cause. More data is needed.

[01:11:49]Specifically, more spectroscopic data at higher red shifts and with more detailed characterization of the confirmed galaxies properties is what will narrow the range of viable explanations.

[01:12:02]That data is coming. The JWST observing program for early galaxies has been growing in scope and ambition since the telescope's first year of operations.

[01:12:13]The JADES program, the JWST advanced deep extragalactic survey, is conducting the deepest systematic spectroscopic survey of high redshift galaxies ever performed, targeting known phototric galaxy candidates in the Hubble Ultra Deep Field and surrounding areas with nearspec spectroscopy that will confirm red shifts, measure chemical abundances, and characterize star formation histories for hundreds of galaxies at red shifts above five. The first Jade's data releases have already produced numerous results and the program is expected to continue for multiple JWST observing cycles. The cosmic evolution early release science program, the primer survey, the uncover survey, multiple independent programs are simultaneously building up the statistical sample of high redshift galaxies with both phototric and spectroscopic characterization.

[01:13:09]The field is moving fast in the way that fields move when a new instrument provides a qualitatively different observational capability than anything that came before. There is one last thing I want to say about the JWST early galaxy problem and it concerns what it would mean to discover that the standard cosmological model genuinely needs modification not just astrophysical revision but cosmological revision. The lambda CDM model rests on a set of deep theoretical foundations.

[01:13:44]Einstein's general relativity as the theory of gravity. The cosmological principle that the universe is homogeneous and isotropic on large scales, inflation as the mechanism that produced the primordial density fluctuations, and the specific particle physics content of the universe, including cold dark matter and a cosmological constant.

[01:14:08]To modify the model at the cosmological level means changing something about these foundations. either modifying gravity on cosmological scales or relaxing the cosmological principle or revising the inflation model or changing the nature of dark matter or dark energy. Okay.