Over-massive Black Holes

• Over-massive black holes (OMBHs) are central black holes whose masses significantly exceed canonical host-galaxy scaling relations, indicating anomalous growth histories.
• Environmental processes, from tidal stripping in dwarfs to dry mergers in clusters, drive these OMBHs beyond typical mass predictions.
• Diverse formation channels—from heavy seeding and rapid accretion at high redshift to slow stellar buildup in satellites—explain the wide range of observed OMBH properties.

Searching arXiv for papers on over-massive black holes, black hole–host scaling relations, and ultramassive black holes. Over-massive black holes (OMBHs) are central massive black holes whose masses lie well above the black-hole–host-galaxy scaling relations usually taken as the baseline for black-hole/galaxy coevolution. In current usage, the class spans several observational regimes: dwarf and satellite galaxies whose central black holes are tens of times above the local $M_{\rm BH}$–$M_\star$ relation, high-redshift AGN and “overmassive black hole galaxies” whose $M_{\rm BH}/M_\star$ ratios exceed local values by factors of several to two orders of magnitude, and ultramassive black holes in brightest cluster galaxies that sit above canonical $M_{\rm BH}$–$\sigma_\ast$ expectations while approaching the empirical high-mass ceiling near $10^{10}\,M_\odot$ (Weller et al., 2023, Beckmann et al., 28 Apr 2026, Nicola et al., 3 Dec 2025).

1. Definition and reference relations

OMBHs are defined relative to empirical black-hole scaling relations. A commonly used local rule of thumb is that the central massive black hole mass is typically about $M_{\rm BH}/M_\ast \sim 10^{-3}=0.1\%$, while bulge-dominated systems are often quoted at $0.2\%-0.5\%$ of the bulge mass; in this framework, an over-massive system is one lying significantly above the mean $M_{\rm BH}$–$\sigma_\ast$, $M_\star$0–$M_\star$1, or $M_\star$2–$M_\star$3 relation after accounting for intrinsic scatter and measurement uncertainties, with a pragmatic threshold often taken as $M_\star$4–$M_\star$5 dex above the median relation (Beckmann et al., 28 Apr 2026, Trakhtenbrot et al., 2015).

In simulation work focused on Leo I, over-massiveness is quantified as the vertical logarithmic offset $M_\star$6, where $M_\star$7 is the black-hole mass predicted by a chosen $M_\star$8–$M_\star$9 relation. Positive offsets denote over-massive systems; the literature then distinguishes strong outliers such as objects $M_{\rm BH}/M_\star$0 above the relation and Leo I-like systems at $M_{\rm BH}/M_\star$1–$M_{\rm BH}/M_\star$2 above it (Weller et al., 2023).

At the high-mass end, related terminology partly overlaps with OMBH usage. Natarajan and Treister define ultra-massive black holes as $M_{\rm BH}/M_\star$3 and argue for a likely upper limit of order $M_{\rm BH}/M_\star$4, while recent brightest-cluster-galaxy work adopts $M_{\rm BH}/M_\star$5 as the operational UMBH regime; these systems are over-massive primarily relative to the canonical $M_{\rm BH}/M_\star$6–$M_{\rm BH}/M_\star$7 relation rather than to total stellar mass alone [(0808.2813); (Nicola et al., 3 Dec 2025)].

2. Nearby dwarfs, satellites, and the low-redshift environmental route

The clearest nearby dwarf-galaxy example is Leo I. Its dynamical black-hole mass is $M_{\rm BH}/M_\star$8 and its stellar mass is $M_{\rm BH}/M_\star$9, placing it $M_{\rm BH}$0–$M_{\rm BH}$1 times above the standard local $M_{\rm BH}$2–$M_{\rm BH}$3 relation. In ASTRID, at the stellar mass of Leo I, about $M_{\rm BH}$4 of galaxies already above the relation are $M_{\rm BH}$5 above it, but Leo I-like extremes are rare: $M_{\rm BH}$6 in the Leo I mass bin, corresponding to $M_{\rm BH}$7 of all over-massive systems (Weller et al., 2023).

The same study isolates a distinct low-redshift formation channel. In TNG50 galaxies with $M_{\rm BH}$8, over-massive hosts followed for $M_{\rm BH}$9 show slower stellar assembly, steeper gas-fraction decline, lower star-formation rates at $\sigma_\ast$0, and only marginally faster SMBH growth than under-massive systems. Median stellar mass growth is $\sigma_\ast$1 for over-massive systems versus $\sigma_\ast$2 for under-massive systems; median gas fraction declines by $\sigma_\ast$3 versus $\sigma_\ast$4; median black-hole growth is $\sigma_\ast$5 versus $\sigma_\ast$6. Major black-hole mergers are not the main driver: only one central SMBH in the sample experienced a major BH merger over $\sigma_\ast$7 Gyr, and over-massive and under-massive systems have similar major-galaxy-merger counts. Environment is decisive for the strongest low-$\sigma_\ast$8 outliers: $\sigma_\ast$9 of satellite systems with at least one massive neighbor are over-massive, $10^{10}\,M_\odot$0 with $10^{10}\,M_\odot$1 massive neighbors are over-massive, $10^{10}\,M_\odot$2 of galaxies with $10^{10}\,M_\odot$3 are over-massive, and $10^{10}\,M_\odot$4 with $10^{10}\,M_\odot$5 are over-massive. Direct ASTRID analogs of Leo I lose $10^{10}\,M_\odot$6–$10^{10}\,M_\odot$7 of their stellar mass between $10^{10}\,M_\odot$8 and $10^{10}\,M_\odot$9 while their central SMBHs grow only slightly, showing how satellite infall and tidal stripping can drive a system far above the canonical relation (Weller et al., 2023).

3. Accreting OMBHs at intermediate redshift and cosmic noon

OMBHs are not confined to dwarfs or to the early universe. An eROSITA hard X-ray survey identified 200 quasars and, with SDSS spectroscopy plus UV-to-IR galaxy–quasar decomposition, securely isolated a sample of OMBHs with $M_{\rm BH}/M_\ast \sim 10^{-3}=0.1\%$0, explicitly defined as ten times above local galaxy scaling relations. The survey yielded a high space density of at least $M_{\rm BH}/M_\ast \sim 10^{-3}=0.1\%$1 near cosmic noon and was interpreted as evidence for an accretion channel disconnected from the stellar population. The same analysis argued that these sources may have undergone exponential accretion spurts lasting about a billion years, and that current galaxy-evolution models do not include the relevant channel (Buchner et al., 23 Mar 2026).

A particularly extreme high-$M_{\rm BH}/M_\ast \sim 10^{-3}=0.1\%$2 accreting case is CID–947 at $M_{\rm BH}/M_\ast \sim 10^{-3}=0.1\%$3. Its black-hole mass from broad H$M_{\rm BH}/M_\ast \sim 10^{-3}=0.1\%$4 is $M_{\rm BH}/M_\ast \sim 10^{-3}=0.1\%$5, its conservative stellar mass estimate is $M_{\rm BH}/M_\ast \sim 10^{-3}=0.1\%$6, and its ratio is $M_{\rm BH}/M_\ast \sim 10^{-3}=0.1\%$7, or about $M_{\rm BH}/M_\ast \sim 10^{-3}=0.1\%$8. This places it at least an order of magnitude, and more likely by a factor $M_{\rm BH}/M_\ast \sim 10^{-3}=0.1\%$9, above local high-mass relations. Yet the host is a main-sequence star-forming galaxy at its epoch, with $0.2\%-0.5\%$0, while the AGN currently accretes at only $0.2\%-0.5\%$1–$0.2\%-0.5\%$2. The combination of a very massive BH, low current Eddington ratio, and a BAL outflow with $0.2\%-0.5\%$3 implies a substantially more active earlier phase of black-hole growth, while the host was still building its stellar mass (Trakhtenbrot et al., 2015).

These intermediate-redshift systems are important because they demonstrate that very large $0.2\%-0.5\%$4 ratios are not unique to seed-formation physics. In the eROSITA interpretation, evolved galaxies can also host OMBHs, so over-massiveness by itself does not isolate a primordial heavy-seed channel (Buchner et al., 23 Mar 2026).

4. Cosmic dawn, heavy seeds, and the high-redshift OBG regime

At high redshift, over-massive black holes are commonly discussed through “overmassive black hole galaxies” (OBGs). In this regime the diagnostic quantity is again $0.2\%-0.5\%$5, but the offsets are larger: local galaxies typically have $0.2\%-0.5\%$6, whereas JWST AGN at $0.2\%-0.5\%$7 often appear higher by factors of several to an order of magnitude, and the most extreme $0.2\%-0.5\%$8 systems reach $0.2\%-0.5\%$9–$M_{\rm BH}$0 (Beckmann et al., 28 Apr 2026, Latif et al., 30 Mar 2026).

One explicit pathway is direct-collapse seeding. In a radiation-hydrodynamic cosmological simulation, a $M_{\rm BH}$1 DCBH forms at $M_{\rm BH}$2 in a $M_{\rm BH}$3 halo, grows at an average $M_{\rm BH}$4, and reaches $M_{\rm BH}$5 by $M_{\rm BH}$6. Its host reaches $M_{\rm BH}$7, $M_{\rm BH}$8, and $M_{\rm BH}$9, giving $\sigma_\ast$0. The high ratio arises from a sequence in which X-ray feedback from the DCBH suppresses star formation for about $\sigma_\ast$1 Myr and later Pop III supernovae violently blow out metals and delay efficient Pop II star formation for $\sigma_\ast$2 Myr; during that interval the BH continues to grow while the stellar component lags (Latif et al., 30 Mar 2026).

A second recent route emphasizes proto-cluster environments and heavier seeds. Fully cosmological radiation-hydrodynamic simulations find heavy seeds of order $\sigma_\ast$3 forming under intense LW irradiation in halos with $\sigma_\ast$4, followed by short super-Eddington episodes that drive growth to $\sigma_\ast$5 by $\sigma_\ast$6. The same calculation links the dense, optically thick early accretion phase to “little red dots,” with H$\sigma_\ast$7 luminosities up to $\sigma_\ast$8, Thomson depths of order $\sigma_\ast$9, and a predicted number density $M_\star$00 for bright massive-BH hosts. In parallel, an analytic halo-driven model proposes that OBHs are a transient phase in which halo gravity drives rapid early accretion before a transition to a hot, pressure-supported halo suppresses BH growth and steers the system back toward the local $M_\star$01–$M_\star$02 relation; in that framework, LRDs are a likely observational manifestation of the rapid halo-driven phase (Chon et al., 8 Jan 2026, Sharma et al., 15 May 2026).

The high-$M_\star$03 literature therefore does not support a single origin. Heavy seeds, DCBH birth, super-Eddington growth, and halo-driven cold accretion are all presented as viable mechanisms for producing OMBHs at cosmic dawn, whereas the observational common denominator is a BH that has run ahead of the stellar body of its host (Beckmann et al., 28 Apr 2026).

5. Ultramassive black holes, brightest cluster galaxies, and the high-mass ceiling

At the opposite extreme in host mass, OMBHs appear as ultramassive black holes in giant ellipticals and brightest cluster galaxies. Natarajan and Treister argue that ultra-massive black holes with $M_\star$04 should exist naturally as the high-mass tail of the population, but that feedback-regulated growth imposes an upper limit of order $M_\star$05. Independent observational reviews likewise note that empirical BH mass functions show a sharp decline above $M_\star$06, consistent with a maximum BH mass of order $M_\star$07 [(0808.2813); (Vestergaard et al., 2023)].

Recent triaxial Schwarzschild modelling of 16 BCGs doubled the directly measured $M_\star$08 sample by discovering 8 new UMBHs. In this regime, BCGs are clear outliers in the canonical $M_\star$09–$M_\star$10 relation, whereas core structural quantities become the best predictors of black-hole mass. The measured relation is

$M_\star$11

with intrinsic scatter $M_\star$12 dex, and the sphere-of-influence radius obeys

$M_\star$13

so that $M_\star$14 across the sample. These tight core-size, sphere-of-influence, and core-density correlations are interpreted as strong support for black-hole binary scouring in dry-merger-built core ellipticals, and they show that at the highest masses “over-massiveness” is a statement about the breakdown of global $M_\star$15-based scaling rather than an absence of coevolution altogether (Nicola et al., 3 Dec 2025).

This high-mass regime is therefore distinctive. OMBHs in BCGs are not primarily small-host anomalies; they are the endpoints of dry-merger assembly, core excavation, and feedback-limited growth near the empirical ceiling [(0808.2813); (Nicola et al., 3 Dec 2025)].

6. Physical pathways, controversies, and methodological limits

The literature converges on the conclusion that OMBHs are not a single phenomenon. At low redshift in dwarfs and satellites, the dominant mechanisms are suppressed stellar buildup, gas depletion, and environmental processing, especially tidal stripping in overdense regions. At high redshift, heavy seeds and rapid accretion dominate the discussion. At cluster scales, dry mergers and black-hole binary scouring govern the most massive examples. This redshift-dependent split is stated explicitly in the Leo I simulation study: high-$M_\star$16 over-massive systems are signatures of heavy black-hole seeds, whereas low-$M_\star$17 over-massive systems result from complex environmental interactions (Weller et al., 2023).

The main controversies concern whether apparent over-massiveness is physical or observational. Measurement systematics are substantial: even direct dynamical masses carry uncertainties up to $M_\star$18 dex, while single-epoch virial estimates can be uncertain by up to $M_\star$19 dex. Flux-limited AGN samples preferentially select high accretion rates and high $M_\star$20 at fixed host mass; spatial-resolution limits in dynamical studies favor large spheres of influence; and at low masses simulations often flatten near the seed mass, making both the reference relation and the inferred tail sensitive to seeding prescriptions and resolution. In addition, many simulations reproduce the mean $M_\star$21–$M_\star$22 and $M_\star$23–$M_\star$24 relations but under-predict the observed scatter, especially at $M_\star$25, which can translate into too few genuine OMBHs (Beckmann et al., 28 Apr 2026, Weller et al., 2023).

A further complication is that different OMBH populations are over-massive relative to different baselines. Leo I is over-massive relative to $M_\star$26–$M_\star$27; BCGs are especially over-massive relative to $M_\star$28–$M_\star$29; some disc-dominated galaxies are over-massive relative to bulge mass rather than total stellar mass; and high-$M_\star$30 JWST AGN can look less extreme when compared with local quiescent-galaxy relations than with local AGN samples (Beckmann et al., 28 Apr 2026).

The most robust synthesis is therefore classificatory rather than monolithic. OMBHs are best understood as a family of outliers above standard black-hole–host relations whose physical origin changes with mass scale, environment, and cosmic epoch: heavy-seed and rapid-accretion signatures at cosmic dawn, differential growth and stripping fossils at low redshift, and ultramassive dry-merger remnants in the centers of the most massive galaxies (Weller et al., 2023, Nicola et al., 3 Dec 2025, Buchner et al., 23 Mar 2026).