For years the centre of the Milky Way has been treated as a settled location. The motions of the bright S cluster of stars near Sgr A* were considered the final word. Their tight, fast, repeated swings around a compact object were presented as direct evidence for a four million solar mass black hole. That conclusion has been repeated so often that it hardened into routine knowledge. The orbits matched predictions, the speeds were extreme, and the star known as S2 delivered clean data through two full passages near periapse. Nothing about the central region appeared uncertain. A new line of work now challenges this view with a different framework that matches the same motions without requiring a black hole at all. Instead of an event horizon, the gravitational pull could come from a dense core of dark matter compressed into a small radius and shaped by the properties of the particles that compose it.
This model uses fermions with masses tens to hundreds of times higher than the electron. These particles can form compact, self gravitating structures when placed under the conditions found in the inner Galaxy. When such a system reaches equilibrium, the interior becomes supported by degeneracy pressure while the outer regions spread into a diluted halo. The result is a single configuration that behaves like a central compact body while also producing an extended dark matter profile at large radii. The rotation curve of the Milky Way has been used to test this scenario on large scales. The new work pushes the test inward by building full general relativistic orbits for the stars that pass closest to Sgr A*.
Two particle masses are central to this effort. A fermion of 56 keV produces a core that is compact but still noticeably extended. A fermion of 300 keV produces a much denser core that approaches the kind of gravitational field usually associated with a black hole. Both can be tuned so that the mass enclosed inside the orbit of S2 reaches the required four million solar masses. What changes between them is how sharply the density climbs as the radius shrinks. This shape determines the metric around the centre, which in turn sets the exact path that any star must follow.
To test this, the motion of S2 is reconstructed using full geodesics. The star’s orbit is launched from apocentre with energy and angular momentum fixed by the geometry, and the path is traced in curved spacetime. The three dimensional orbit is then rotated onto the sky using classical transformations and compared with the observed positions and velocities recorded over nearly two decades. This process is repeated across a wide range of parameters using a sampler that explores possible combinations and identifies those that match the data within known uncertainties.
The result is striking. The orbits generated by the two fermionic models and the orbit generated by a four million solar mass black hole are almost identical across the entire observational window. The position curves overlap. The radial velocity curves overlap. The periapse distance matches the measured value. The apocentre separation matches as well. Even the argument of periapse, inclination, and orientation of the ascending node are reproduced to within fractions of a degree. Every parameter that can be extracted from the S2 tracking falls within the same narrow range across all three gravitational potentials. No current measurement distinguishes them.
This lack of separation is not the product of weak modelling. It arises because all three gravitational fields converge at the radius where S2 travels. The enclosed mass is fixed by observation. If a dark matter core supplies that mass at the correct radius, its influence on the orbit becomes practically indistinguishable from that of a black hole even though the internal structure is completely different. Only measurements taken much closer to the central object could reveal the differences between these potentials.
The work continues by examining the G objects, a class of dusty, red, compact sources that move on long, eccentric trajectories inside the inner tenth of a parsec. These objects have shorter observational coverage and lower positional precision than S2. Their true nature has been debated for a decade. They were once classified as gas clouds expected to be torn apart, yet they remained intact after their closest approaches. Later studies argued for a stellar core wrapped in gas and dust. Revised orbital fits cleanly separated their line emission from unrelated background sources. These updated positions and velocities are used here to test the gravitational potentials further.
The G objects have incomplete orbits. The segments observed so far cover only a small part of each path. For that reason the central potential cannot be refit for each one. Instead, the parameters determined from S2 are held fixed and only the orbital elements are varied. The trajectories generated for G2, G3, G4, G5, and G6 under all three gravitational fields are compared with the data. The result again shows no clear separation. Some fits favour one potential slightly more than another, but none by a margin that allows a secure identification of the underlying mass distribution. The orbital arcs are simply too short and too distant from the centre to sense the differences between a true event horizon and a dense dark matter core.
A separate test checks how strongly the inner core parameters depend on the shape of the outer halo. The mass distribution at kiloparsec radii has been updated by recent stellar kinematic surveys. When these updated halo masses are used, the entire fit for S2 remains stable. The orbital parameters shift by less than a tenth of a per cent. This demonstrates that the behaviour of S2 is sensitive only to the mass inside its path and not the details of the halo far beyond it. The dark matter core required to match the observed motion is therefore not an artifact of distant mass assignments. It is set by the conditions near the centre.
A further comparison plots the predicted rotation curve from the fermionic core halo assignments against the observed rotation curve of the Milky Way across all radii. The model reproduces the behaviour of the inner bulge, matches the contribution from the disc, and falls in line with the outer halo velocities measured by recent surveys. The circular speeds expected at the radius of S2 and the location of G2 sit naturally on the continuation of this curve. This provides a unified view in which the same dark matter configuration dictates both the large scale structure and the compact central region. In contrast, the black hole model addresses only the central pull and leaves the outer rotation curve to separate dark matter distributions.
A precise point emerges from this entire effort. The centre of the Milky Way can be described either by a black hole or by a dense fermionic core, and every current observation of stellar motion allows both. The gravitational field produced by a compact dark matter configuration is strong enough and structured enough to guide S2 through its full sixteen year orbit without deviating from the recorded positions by any measurable amount. The same holds for the G objects whose motions trace shallower sections of the potential. Independent constraints from the galactic rotation curve also support such configurations. The dark matter core is not an ad hoc replacement for a black hole. It is part of a complete model that spans from sub parsec scales to tens of kiloparsecs.
The question then shifts to what could break this degeneracy. The next apocentre and periapse of S2 after 2026 will provide new opportunities for comparison. The instrumental precision now available can register subtle shifts that were beyond reach during earlier observing campaigns. Stars that travel even closer to the centre, if identified and tracked, would test radii where the differences between a horizon and a dense core become larger. Those measurements do not yet exist in the required form, and until they do the gravitational potential at the centre remains open to more than one configuration.
Another line of investigation uses imaging of the region surrounding Sgr A*. High resolution interferometry has produced a brightness depression at the centre of the emission ring. This feature has been described as a shadow consistent with a black hole. However, detailed work on ray tracing around dense fermionic cores shows that similar depressions can be produced without invoking an event horizon. The size of the depression depends on how compact the core is and how the surrounding plasma emits and absorbs radiation. For fermion masses near 300 keV the predicted diameter matches the scale reported for Sgr A*. The presence of a dark centre in the image therefore does not remove the alternative model.
Taken together, these results place the central mass of the Milky Way in a sharper but more complex position. The traditional picture of a four million solar mass black hole is not contradicted by any new observation, but it is also not uniquely required by the data. A dark matter core built from heavy fermions remains fully consistent with every measurement of stellar orbits and with the global rotation curve of the Galaxy. The motions near Sgr A* provide strong evidence for a compact mass, but they do not specify its internal structure. That structure can vary without altering the gravitational pull felt by nearby stars. The evidence now shows that more than one configuration fits the facts directly. The choice between them will depend on data that reach deeper into the central region and capture the next generation of tight orbits around the compact source.
Source:
The dynamics of S-stars and G-sources orbiting a supermassive compact object made of fermionic dark matter, Monthly Notices of the Royal Astronomical Society, 2026.
Link: https://doi.org/10.1093/mnras/staf1854
