For its first two years on sky, the James Webb Space Telescope kept reporting the same uncomfortable thing. Galaxies that looked too bright, too massive, and too far back in time to fit the standard cosmological model. Headlines about “galaxies that break cosmology” piled up. The community quietly fought about how seriously to take them. Two years of follow-up spectroscopy have now settled part of the question, and the answer reframes the puzzle. A large fraction of the offending objects are not absurdly mature stellar systems. They are early galaxies hosting black holes that are far too big for their age.
The original shock
The trouble started with photometric estimates. When JWST began producing deep images in late 2022, several teams used the colors of compact red sources to estimate redshifts (how far back in time we are looking) and stellar masses. A handful of those sources, at apparent redshifts between 7 and 12, came out with stellar masses around 10 billion solar masses, comparable to today’s Milky Way, only 400 to 700 million years after the Big Bang. In the standard Lambda Cold Dark Matter (LCDM) framework, that much stellar material that early is hard to build. The dark-matter halos to host it should not yet exist in those numbers.
Two things turned out to be going on. First, photometric redshifts based on broad-band colors are noisy. Spectroscopic follow-up with the NIRSpec instrument has confirmed that a real population of early galaxies exists out to and beyond redshift 14, but a chunk of the initial photometric sample was less extreme than the headlines suggested. Second, the light from many of these objects is not coming from stars at all.
Enter the little red dots
By 2023, JWST imaging had identified a peculiar population astronomers now call “little red dots.” They are compact (unresolved, or barely resolved), red in optical light, and far more common at high redshift than anyone had predicted. Estimates from the CEERS and JADES surveys put their density at hundreds of times what extrapolation from later epochs would have implied.
Spectroscopy of these dots over 2024 and 2025 returned a consistent signature: broad emission lines, especially hydrogen-alpha, with widths of 2,000 to 4,000 kilometers per second. Broad lines like that come from gas whipping around a central black hole at high speed. They are the signature of an active galactic nucleus.
The implication is that the “too much light” from many compact early sources is not the steady glow of billions of stars, but the brilliant point-source emission of a quasar embedded in a much smaller host. Once you subtract the AGN contribution, the inferred stellar mass drops, often by a factor of ten or more. That alone takes pressure off the cosmological model for the worst-offending objects.
The new problem the dots created
It also creates a different problem. The black holes themselves, weighed by line widths, are too big. Several little red dots host black holes of 10 to 100 million solar masses around 500 million years after the Big Bang. That is more black-hole mass than the standard formation channels (collapse of the first generation of massive stars, then growth at the Eddington-limited accretion rate) can produce in the time available.
The cleanest single example is UHZ1, a galaxy at spectroscopic redshift 10.3 imaged by JWST and detected in X-rays by Chandra. The X-ray luminosity implies a black hole of at least 10 million solar masses already in place 470 million years after the Big Bang. The host galaxy’s stellar mass is roughly comparable to the black hole’s, where in the present-day universe a black hole is typically 0.1 percent of its host’s mass. UHZ1 is therefore a black hole that is roughly a thousand times too heavy for its galaxy, sitting at an epoch where conventional seeds should not yet have grown to that size.
The leading explanation, gathering observational support over the last year, is heavy seeding. Some of the earliest black holes did not start as stellar remnants of a few hundred solar masses. They started as “direct-collapse” objects, formed when massive primordial gas clouds skipped the star-formation step entirely and collapsed straight into a 10,000 to 100,000 solar-mass black hole. Heavy seeds can grow into the observed early quasars without straining the timeline. They are theoretically allowed under specific conditions in the very early universe, and JWST is the first instrument capable of seeing them.
What remains genuinely strange
Removing AGN flux from the budget does not erase the early-galaxy abundance problem entirely. Even the spectroscopically confirmed star-forming galaxies at redshift 12 and beyond, such as JADES-GS-z14-0 at z = 14.32 (light from about 290 million years after the Big Bang), are brighter and more numerous than pre-launch models predicted. Models that solve this without breaking cosmology generally invoke either bursty, very efficient star formation in the earliest halos, or a “top-heavy” initial mass function in which the first stellar populations skewed toward massive, luminous stars that fade quickly.
Both of those are plausible. Neither has been pinned down. The next year of NIRSpec and MIRI spectroscopy, plus the deeper Chandra and forthcoming Athena X-ray follow-ups on the most extreme little red dots, should narrow the options.
Why this matters
The early-galaxy puzzle never threatened cosmology itself the way the headlines sometimes suggested. The Hubble constant tension is a bigger threat to LCDM. What this puzzle threatens, and what it is now starting to reshape, is the story of how the first generation of structure assembled. The picture coming into focus is one where, very early, gas finds shortcuts to making black holes that the local universe does not offer. Galaxies and their central black holes are not co-evolving smoothly all the way back. Near the beginning, the black holes get a head start.
JWST was designed to see this epoch. The early data was confusing. The follow-up data is now resolving the confusion in an interesting direction: not by erasing the surprise, but by relocating it from the stars to the holes.