A single 8,200 M⊙ IMBH (Häberle 2024) and a 3,000 M⊙ dark cluster of stellar remnants (Bañares 2025) fit the same kinematics. Five of six observational programmes admit both. The sixth — astrometric microlensing — does not. That's what makes Gaia DR4 (Dec 2026) the experiment of the decade for ω Cen.
Each scenario pre-loads all six tools with the parameters that scenario predicts. Walk the chain to see which observations distinguish them — and which don't.
Sixteen years of measurements, all overlaid on a single log-mass axis. Häberle 2024 requires M > 8,200 M⊙; Bañares 2025 requires M < 6,000 M⊙. The two bounds do not overlap. One of these results is wrong, or both methods have a hidden systematic.
ω Cen's observed central velocity dispersion is σ ≈ 19–22 km/s, which is what the virial theorem demands from a cluster of its dynamical mass. The cluster's σ is set by the bulk stellar mass, not by what's at the centre. The Gultekin M-σ extrapolation gives a soft "predicted" M_BH from σ alone — but its scatter is ≈ 0.4 dex, big enough to admit anything from a stellar-mass remnant to a 10⁵ M⊙ SMBH.
A Plummer-profile distribution of stellar-mass remnants — a few hundred neutron stars and stellar black holes settling into the core via mass segregation — reproduces the enclosed-mass profile that fast central stars would naively read as a single point mass. The dynamical fit cannot tell the two apart at the central resolution of HST or JWST. This is the central thesis of Bañares-Hernández et al. 2025 (A&A 693:A104).
The two-body relaxation timescale at ω Cen's density and age determines whether heavy remnants segregate into the core fast enough — and whether the runaway-merger channel (Portegies Zwart 2004) can grow an IMBH in situ. The key knob is BH retention: if natal kicks eject most stellar BHs, neither model is feasible; if retention is high, both are. The dark-cluster scenario and the in-situ-IMBH scenario share the same physics, with different end-states.
The TRAPUM pulsars near ω Cen's core constrain the enclosed mass at ≈ 10″. The single-pulsar formal sensitivity reaches ≈ 10³ M⊙ on a decade baseline at μs precision — in principle enough to detect either model — but the dark-cluster and IMBH scenarios both inject >rsim 10³ M⊙ of enclosed mass at that radius. So the same acceleration signal is read as either-or. Pulsar timing's published bound (6,000 M⊙) sits comfortably above both predictions.
A point lens of mass M produces a centroid shift Θ_c that scales as √M and peaks at the Einstein radius θ_E. For an 8,200 M⊙ lens behind ω Cen, Θ_c ≈ hundreds of μas — well above Gaia DR4's 30 μas floor. For a 10 M⊙ stellar remnant in a Bañares dark cluster, Θ_c ≈ 0.4 μas — far below any reachable threshold this decade. Only a point mass produces a detectable centroid shift; a distributed dark cluster does not.
The IMBH question has been stuck since Noyola 2008 because every dynamical observable — bulk σ, central acceleration, accretion luminosity, pulsar timing — depends only on the enclosed mass, not on whether that mass is concentrated at a point or spread across the central parsec. The two scenarios live in different mass distributions but the same Newtonian potential at the resolution we can reach.
Gravitational lensing is the exception. It is sensitive to the lens's compactness, not just its mass. A point lens deflects light around an Einstein ring; a distributed lens of equal total mass deflects it almost not at all because each individual remnant has a tiny Einstein radius. The signal is &sqrt;M for a point and ≈ 0 for the cluster — a binary decision dressed in continuous numbers.
Gaia DR4 is scheduled for late 2026 and reaches ≈ 30 μas for bright bulge sources; Roman launches into operations in 2027 with ≈ 10 μas. The predicted IMBH astrometric centroid-shift amplitude is ~776 μas (Roman-DETECTABLE per the tool) — well above both instruments' noise floors. Either we will have a positive detection within 18 months, or we will have a clean negative result — and either way the decade-long ambiguity ends.
For the broader observational roadmap on every claim made by this site, see the Falsification & Observational Roadmap; for the Fermi-pillar companion to this demo, see Demo O — Five Ways to Look for ET.