JWST observations have unveiled faint active galactic nuclei (AGNs) at high redshift that provide insights into the formation of supermassive black holes (SMBHs). However, disentangling their stellar from AGN light is challenging. Here, we use an empirical approach to infer the average stellar mass of five faint broad-line (BL) Hα emitters at z = 4–5 with BH masses ≈6 × 10⁶ M _⊙, with a method independent of their spectral energy distribution (SED). We use the deep JWST/NIRcam grism survey “All the Little Things” to measure the overdensities around BL-Hα emitters and around a spectroscopic reference sample of ∼300 galaxies. In our reference sample, we find that megaparsec-scale overdensity correlates with stellar mass. Their large-scale environments suggest that BL-Hα emitters are hosted by galaxies with stellar masses ≈5 × 10⁷ M _⊙, ≈40 times lower than those inferred from galaxy-only SED fits. Adding measurements around more luminous z ≈ 6 AGNs, we find tentative correlations between line width, BH mass, and the overdensity, suggestive of a steep BH to halo mass relation. The main implications are (1) when BH masses are taken at face value, we confirm extremely high BH to stellar mass ratios of ≈10%, (2) the galaxies of low stellar mass that host growing SMBHs are in tension with typical hydrodynamical simulations, except those without feedback, (3) a 1% duty cycle implied by the host mass hints at super-Eddington accretion, (4) the masses are at odds with an interpretation of the line broadening in terms of high stellar density, (5) our results imply a luminosity-dependent diversity of galaxy masses, environments, and SEDs among AGN samples.

The population of the little red dots (LRDs) may represent a key phase of supermassive black hole (SMBH) growth. A cocoon of dense excited gas is emerging as a key component to explain the most striking properties of LRDs, such as strong Balmer breaks and Balmer absorption, as well as the weak IR emission. To dissect the structure of LRDs, we analyzed new deep JWST/NIRSpec PRISM and G395H spectra of FRESCO-GN-9771, one of the most luminous known LRDs at z = 5.5. These spectra reveal a strong Balmer break, broad Balmer lines, and very narrow [O III ] emission. We revealed a forest of optical [Fe II ] lines, which we argue are emerging from a dense ( n _H = 10 ^9 − 10 cm ⁻³ ) warm layer with electron temperature T _e ≈ 7000 K. The broad wings of H α and H β have an exponential profile due to electron scattering in this same layer. The high H α : H β : H γ flux ratio of ≈10.4 : 1 : 0.14 is an indicator of collisional excitation and resonant scattering dominating the Balmer line emission. A narrow H γ component, unseen in the other two Balmer lines due to outshining by the broad components, could trace the ISM of a normal host galaxy with a star formation rate of ∼5 M _⊙ yr ⁻¹ . The warm layer is mostly opaque to Balmer transitions, producing a characteristic P Cygni profile in the line centers suggesting outflowing motions. This same layer is responsible for shaping the Balmer break. The broadband spectrum can be reasonably matched by a simple photoionized slab model that dominates the λ > 1500 Å continuum and a low-mass (∼10 ⁸ M _⊙ ) galaxy that could explain the narrow [O III ], with only a subdominant contribution to the UV continuum. Our findings indicate that Balmer lines are not directly tracing the gas kinematics near the SMBH and that the BH mass scale is likely much lower than virial indicators suggest.

The physical processes that led to the formation of billion-solar-mass black holes within the first 700 million years of cosmic time, a period known as cosmic dawn, remain a puzzle ¹ . Several theoretical scenarios have been proposed to seed and rapidly grow black holes ^2–4 , but direct observations of these mechanisms remain elusive. Here we present a source 660 million years after the Big Bang that exhibits singular properties: among the largest hydrogen Balmer breaks reported at any redshift, broad multi-peaked Hβ emission, and Balmer line absorption in several transitions. We model this source as an enshrouded black hole in which the Balmer break and absorption features are a result of extremely dense, turbulent gas forming a dust-free envelope around a supermassive black hole ^5,6 . This source may provide evidence of an early black hole embedded in dense gas—a theoretical configuration proposed to rapidly grow black holes by super-Eddington accretion ^7,8 . Radiation from the black hole seems to dominate almost all observed light, leaving limited room for contribution from its host galaxy. If the source merged with its brighter neighbour, it would resemble the recently discovered ‘little red dots’ with perplexing spectral energy distributions ^9–11 . The redness of the black hole is due to gas, not dust ^12,13 , and scattering, not kinematics, gives rise to the complex line shapes and luminosities—black hole masses of these sources may therefore be overestimated by orders of magnitude.

The increasingly neutral intergalactic gas at z > 6 impacts the Lyman-α (Lyα) flux observed from galaxies. One luminous galaxy, COLA1, stands out because of its unique double-peaked Lyα line at z = 6.6, unseen in any simulation of reionization. Here, we present JWST/NIRCam wide-field slitless spectroscopy in a 21 arcmin² field centered on COLA1. We find 141 galaxies spectroscopically selected through the [O III] doublet at 5.35 < z < 6.95, with 40 of these sources showing Hβ. For COLA1, we additionally detect [O III]₄₃₆₃ as well as Hγ. We measure a systemic redshift of z = 6.5917 for COLA1, confirming the classical double-peak nature of the Lyα profile. This implies that it resides in a highly ionized bubble and that it is leaking ionizing photons with a high escape fraction of f_esc(LyC) = 20–50%, making it a prime laboratory to study Lyman continuum escape in the Epoch of Reionization. COLA1 shows all the signs of a prolific ionizer with a Lyα escape fraction of 81 ± 5%, Balmer decrement indicating no dust, a steep UV slope (β_UV = −3.2 ± 0.4), and a star-formation surface density ≳10× that of typical galaxies at similar redshift. We detect five galaxies in COLA1’s close environment (Δz < 0.02). Exploiting the high spectroscopic completeness inherent to grism surveys, and using mock simulations that fully mimic the selection function, we show that the number of detected companions is very typical for a normal similarly UV-bright (M_UV ∼ −21.3) galaxy – that is, the ionized bubble around COLA1 is unlikely to be due to an excessively large over-density. Instead, the measured ionizing properties suggest that COLA1 by itself might be powering the bubble required to explain its double-peaked Lyα profile (R_ion ≈ 0.7 pMpc), with only minor contributions from detected neighbors (−19.5 ≲ M_UV ≲ −17.5).

We compile a sample of 83 little red dots (LRDs) with JWST imaging and find that a substantial fraction (∼43%, rising to ≳80% for the most luminous LRDs) host one or more spatially offset, UV-bright companions at projected separations of 0.5 kpc ≲ d ≲ 5 kpc, with median 〈 d 〉 = 1.0 kpc. This fraction is even higher when smaller spatial scales are probed at high signal-to-noise ratio: the two most strongly lensed LRDs, A383-LRD1 and the newly discovered A68-LRD1, both have UV-bright companions at separations of only d ∼ 0.3 kpc, below the resolution limit of most unlensed JWST samples. We explore whether these ubiquitous red/blue configurations may be physically linked to the formation of LRDs, in analogy with the “synchronized pair” scenario originally proposed for direct-collapse black hole formation. In this picture, UV radiation from the companions, with typically modest stellar masses ( M _∗ ∼ 10 ⁸ −10 ⁹ M _⊙ ), suppresses molecular hydrogen cooling in nearby gas, allowing nearly isothermal collapse and the formation of extremely compact objects, such as massive black holes, supermassive stars, or quasi-stars. Using component-resolved photometry and spectral energy distribution modeling, we infer Lyman–Werner radiation fields of J _21,LW ∼ 10 ^2.5 –10 ⁵ at the locations of the red components, comparable to those required in direct-collapse models, suggesting that the necessary photodissociation conditions are realized in many LRD systems. This framework provides a simple and self-consistent explanation for the extreme compactness and distinctive spectral properties of LRDs and links long-standing theoretical models for early compact object formation directly to a population now observed with JWST in the early Universe.

We construct and characterize a sample of 146 little red dots (LRDs) across ${2.0{\lt} z{\lt} 9.3}$, selecting all sources with v-shaped ultraviolet (UV)–optical continua from James Webb Space Telescope (JWST) NIRSpec/PRISM spectra and compact morphologies in NIRCam/F444W imaging. We show that LRD continuum spectra are well described by modified blackbodies across $\sim 0.4-1.0\, \rm{\mu m}$, with typical $T\sim 5000\,$ K ($\lambda _{\rm peak}\sim 0.65\, \rm{\mu m}$) across 2 dex in luminosity, and a tail toward $T\sim 2000\,$ K. LRDs therefore trace a locus in the Hertzsprung–Russell diagram that is analogous to stars on the Hayashi track, strongly supporting the picture that LRDs are active galactic nucleus (AGN) embedded in optically thick dense gas envelopes. Hotter LRDs with $\lambda _{\rm peak}\lt 0.65\, \rm{\mu m}$ typically have strong Balmer breaks, redder UV slopes, and high optical luminosities; other LRDs show weak or no Balmer breaks, and wide variety in $\beta _{\rm UV}$ and $L_{5100}$. Crucially, we demonstrate that the UV–optical continua are strongly linked to the H$\alpha$, H$\beta$, [O iii], and O i line properties. There is a tight linear relation between the H$\alpha$ and optical continuum luminosities, as well as H$\alpha$ and $\rm {O\, {\small I} }_{\ \lambda 8446}$, indicating that Balmer, O i, and optical emission must primarily be powered by the same source. The Balmer decrement increases strongly toward higher $L_{\rm H\alpha }$, $L_{5100}$, and Balmer break strength, providing key evidence for luminosity-dependent effects of collisional (de-)excitation and resonant scattering in the gaseous envelopes. In contrast, we show that [O iii] emission likely originates from star-forming host galaxies, with strong variation in the AGN-to-host ratio among the LRD population. Our work presents an empirical description of the nature and structure of LRDs, defining a new benchmark for ongoing LRD model developments.