JWST Catches Two Post-Quasar Galaxies With Overmassive Black Holes 800 Million Years After the Big Bang

Overview

Astronomers using NASA’s James Webb Space Telescope have identified two galaxies in the early universe where supermassive black holes are vastly more massive than their host galaxies would normally support — and found the most likely explanation in a process that had not previously been observed in detail: the aftermath of a recently extinguished quasar. The findings, published on the arXiv preprint server on May 1, 2026, by a team led by Romain A. Meyer of the University of Geneva, offer the clearest picture yet of what an early universe galaxy looks like in the hours, cosmologically speaking, after its central engine goes quiet.

The two objects, named COLA1 (redshift z = 6.59075 ± 0.00003) and NEPLA4 (redshift z = 6.54422 ± 0.00002), both appear as they existed roughly 800 million years after the Big Bang. The team’s arXiv preprint titles the paper “Life After the Quasar: Overmassive Black Holes and Remnant Ionised Bubbles in and Around Two z~6.6 Galaxies.”

What We Know

The discovery rests on JWST/NIRSpec integral field unit spectroscopy using the G235M/F170LP and G395M/F290LP gratings, supplemented by Subaru, Spitzer, and ground-based data from Keck and the VLT. The spectra of both galaxies show broad hydrogen Balmer emission lines — a definitive signature of an active nucleus — but at far lower luminosity than a typical quasar. According to the arXiv paper, the broad Balmer lines indicate “black hole masses M_BH ≃ 2×10⁸ M_☉, matching that found in faint z∼6−7 quasars.”

Conventional wisdom in galaxy evolution holds that black holes and their host galaxies grow together, maintaining a roughly constant ratio. In the local universe, black holes typically account for about 0.1–0.5 percent of their galaxy’s stellar mass. COLA1 and NEPLA4 shatter that expectation. The team measures a black hole mass of log₁₀M_BH = 8.28 ± 0.05 M_☉ for COLA1, against a stellar mass of log₁₀M_* = 8.93⁺⁰·⁰⁴₋₀.₀₄ M_☉ — yielding a black hole-to-stellar-mass ratio of 0.22, or about one fifth of the galaxy’s total stellar content. NEPLA4 shows a ratio of 0.10. As Phys.org reports, these ratios are “400–800 times greater than nearby universe observations.”

What makes the case even more striking is the star-formation activity. Both galaxies are currently forming stars at a vigorous rate — 58 ± 2 M_☉/yr for COLA1 and 50 ± 1 M_☉/yr for NEPLA4, according to the arXiv paper — yet their stellar populations are extremely young. The paper finds that “≳90% of the stellar mass was only created in the previous ∼50 Myr,” meaning the galaxies have only recently started assembling stars in earnest. Meanwhile, the central black holes appear to have been feeding voraciously for hundreds of millions of years. Phys.org reports that “black holes began feeding at near-maximum rates when the universe was 180–270 million years old,” long before these galaxies’ dominant stellar populations formed.

The current state of the AGN in both objects underscores the transition. Eddington ratios — a measure of how close to the maximum possible rate a black hole is currently accreting — are low: 0.050 ± 0.004 for COLA1 and 0.039 ± 0.004 for NEPLA4. As Phys.org describes it, “black holes are now dormant, feeding at a fraction of their maximum rate, while the galaxies are undergoing an intense burst of star formation.”

The team’s interpretation, drawn from the arXiv paper, is that both systems are “post-quasar galaxies in which AGN feedback has delayed stellar mass assembly.” The idea is that the central black hole, while active as a quasar, radiated enough energy to suppress or disrupt star formation across the galaxy. Once the quasar faded, star formation restarted rapidly. The paper states: “COLA1 and NEPLA4 have recently (≲1 Myr) transitioned from a quasar phase to a lower-luminosity AGN, with star-formation rapidly picking up after being inhibited.”

A key piece of evidence for the recent quasar shutdown comes from an unexpected spectral feature: double-peaked Lyman-alpha emission. This pattern is exceptionally rare. Neutral hydrogen in the intergalactic medium absorbs the blue wing of the Lyman-alpha line almost completely, so detecting both peaks requires a large, cleared bubble of ionized gas surrounding the galaxy. According to the full paper, such profiles “require a large ionised bubble and high photoionisation rate that is consistent with the ionising output of quasars powered by black holes of similar mass.” The bubble sizes inferred are 0.24–0.29 physical megaparsecs in minimum radius. The blue-peak detection effectively tells astronomers that a quasar was active very recently — the paper constrains “the cessation of the last quasar episode to <1 Myr,” a cosmic eyeblink.

The Lyman-alpha peak separations themselves are measurable: 218 ± 3 km/s for COLA1 and 249 ± 7 km/s for NEPLA4, both consistent with the bubble model. The surrounding medium is almost completely ionized, with the full paper measuring “extremely low neutral fraction xHI∼10⁻⁶.”

Beyond the properties of the two individual galaxies, the study carries implications for the broader Epoch of Reionization — the period roughly 150 million to 1 billion years after the Big Bang when ultraviolet radiation from early stars and galaxies ionized the universe’s hydrogen. The authors argue that “episodic quasar activity partially explains the unexpected prevalence of large ionised bubbles deep into the Epoch of Reionisation,” a phenomenon that previous models of galaxy-driven reionization have struggled to account for. They also estimate that “~11% of Lyman-α emitters without apparent galaxy overdensities may be tracing relics of past quasar/AGN activity,” suggesting that post-quasar systems like COLA1 and NEPLA4 are not isolated curiosities but a detectable population.

What We Don’t Know

The study is a preprint and has not yet undergone peer review. The black hole masses are estimated from broad emission lines under assumptions about the geometry of the emitting region — a method with known systematic uncertainties of roughly a factor of two to three in either direction. The host galaxy stellar masses carry their own uncertainties from population synthesis modeling.

The mechanism that seeded these black holes to reach 2×10⁸ solar masses by the time the universe was only a few hundred million years old remains unresolved. The paper does not identify the original seed — whether from direct collapse of pristine gas clouds, runaway stellar collisions in dense star clusters, or the remnants of the first-generation Population III stars. The post-quasar picture explains why the galaxies are now underdeveloped relative to their black holes, but not how the black holes became so massive so quickly in the first place.

It is also unclear how representative COLA1 and NEPLA4 are. Both were selected from a sample of known double-peaked Lyman-alpha emitters, which is itself a rare subclass. Whether the post-quasar pathway is common among all early overmassive black hole hosts, or whether double-peaked emitters are a biased subset, will require systematic searches across wider fields.

Analysis

The COLA1 and NEPLA4 results add a new observational category to the growing JWST inventory of puzzling early-universe objects. Since 2022, JWST has repeatedly found black holes in high-redshift galaxies that are too massive for their hosts — a pattern collectively referred to as overmassive black holes. What this study adds is a temporal narrative: rather than a static snapshot of an unusual mass ratio, it presents a before-and-after account, with the ionized bubbles serving as a fossil record of the quasar episode that just ended.

The AGN feedback framework invoked here — where black hole energy injection suppresses star formation, then recedes to allow a starburst — has been a staple of galaxy evolution models for two decades. But it has been difficult to observe in real time at high redshift. Finding two galaxies apparently caught in the immediate post-quasar phase, within 1 Myr of the central engine dimming, is either a lucky coincidence or evidence that such transitions are more common, and shorter-lived, than models have assumed. The authors’ estimate that ~11% of double-peaked Lyman-alpha emitters belong to this class points toward a statistically relevant channel rather than a rare accident.

For the broader reionization picture, the contribution of quasar-carved ionized bubbles to the patchwork of reionization is a live debate. Galaxy-driven models require a high escape fraction of ionizing photons, yet direct measurements remain difficult. If short-lived quasar episodes carved significant bubbles and left post-quasar galaxies behind them — with the bubbles persisting long enough to be detectable — then the contribution of nuclear activity to reionization may have been systematically underestimated. COLA1 and NEPLA4 are the clearest examples yet of that process caught in the act.