Astronomers working with Hubble Space Telescope and Euclid data have identified an object in the Perseus Cluster that appears to be composed almost entirely of dark matter, with only the faintest trace of visible stars. The object, designated Candidate Dark Galaxy-2, or CDG-2, was detected through extremely faint diffuse emission surrounding four globular clusters, and it has been described as potentially one of the most heavily dark-matter-dominated galaxies ever discovered. The finding arrives at a time when separate research teams have been reporting galaxies at the opposite extreme, ones that seem to lack dark matter altogether, sharpening a fundamental puzzle about how galaxies form and survive inside dense cluster environments.
Why CDG-2 challenges galaxy formation models right now
CDG-2 sits near the Perseus Cluster, one of the most massive galaxy clusters in the nearby universe. That location matters because clusters like Perseus are gravitationally violent places. Tidal forces strip stars and gas from smaller galaxies as they orbit through the cluster core. A galaxy that retains an enormous dark-matter halo while holding onto almost no visible stellar mass raises a pointed question: how did it form so few stars in the first place, and how has its dark-matter envelope survived the cluster’s gravitational churning?
The tension sharpens when CDG-2 is placed alongside galaxies like NGC 1052-DF2, the dwarf galaxy first reported as lacking dark matter based on the radial velocities of its globular-cluster-like tracers. A third dark-matter-deficient galaxy in the same NGC 1052 field was described in a 2026 paper, reinforcing the idea that extreme outliers on both ends of the dark-matter spectrum are real and not statistical flukes. CDG-2 occupies the opposite pole. If its globular-cluster population traces a massive, early-formed dark-matter halo that survived Perseus-cluster tidal processing, then its globular cluster specific frequency, the number of clusters per unit galaxy luminosity, should far exceed that of both dark-matter-deficient dwarfs and typical ultra-diffuse galaxies at comparable brightness. That prediction can be tested with deeper photometry from Euclid or the Nancy Grace Roman Space Telescope before spectroscopic follow-up even begins.
Stacked Hubble imaging and Euclid confirmation behind CDG-2
The detection pipeline that produced CDG-2 relies on a statistical technique developed in earlier Perseus Cluster work. Researchers used a Log-Gaussian Cox Process to identify over-densities of globular clusters that signal the presence of ultra-diffuse galaxies too faint to detect through conventional surface-brightness searches. That method treats clusters of old, bright star systems as signposts for otherwise invisible host galaxies, allowing astronomers to search for diffuse stellar halos indirectly rather than by their light alone.
For CDG-2 specifically, the team stacked Hubble ACS exposures to pull extremely faint diffuse emission out of the noise around four globular clusters. Euclid imaging then provided independent confirmation of the signal. The combined dataset, described in a recent preprint accepted by the Astrophysical Journal Letters, yielded quantitative luminosity estimates placing CDG-2’s stellar mass at extraordinarily low levels relative to its inferred halo mass. In a separate communication, NASA highlighted the object as potentially “among the most heavily dark matter-dominated galaxies ever discovered,” framing it as the strongest candidate to date for a galaxy made almost entirely of dark matter.
The use of two independent instruments, Hubble’s Advanced Camera for Surveys and the European Space Agency’s Euclid, adds weight to the claim. Each telescope operates in different wavelength bands and has different noise characteristics, so agreement between the two reduces the chance that the diffuse emission is an artifact of a single detector or data-reduction pipeline. The Euclid confirmation is especially important because its wide-field imaging provides a consistent background against which to measure the faint glow, while Hubble’s higher resolution isolates individual globular clusters that act as tracers of the underlying halo.
What spectroscopy and deeper photometry still need to settle
The strongest limitation of the CDG-2 result is that all mass inferences remain photometric. No radial velocity or dynamical mass measurements from spectroscopy have been reported for this object. The globular-cluster overdensity method can identify candidate dark galaxies, but confirming that a given candidate truly sits inside a massive dark-matter halo requires measuring how fast its tracers move. Without those velocities, the halo mass is inferred from the number and spatial distribution of globular clusters rather than from direct gravitational evidence.
The detection pipeline itself was validated in the 2022 methodology paper on Perseus ultra-diffuse galaxies, but no new end-to-end false-positive analysis specific to the CDG-2 field has been published alongside the current result. Selection-function completeness, the question of how many similar objects the pipeline might miss or falsely flag, remains anchored to the earlier calibration rather than an updated assessment tuned to Euclid’s depth and the exact Hubble pointing. That leaves open the possibility that CDG-2 represents an extreme tail of a broader population that current searches are only beginning to glimpse.
No primary data release of the stacked image cutouts or globular-cluster catalogs accompanies the preprint, which limits the ability of independent teams to reproduce the detection before the journal version appears. Direct statements from the lead authors on how CDG-2 fits or strains predictions from the standard cold dark matter model are absent from both the NASA release and the candidate paper; only institutional framing is available so far. For now, the interpretation that CDG-2 is a nearly starless galaxy rests on the internal consistency of the photometry and the robustness of the globular-cluster selection criteria.
The next concrete step to watch
The next concrete step to watch is targeted spectroscopy of the four globular clusters associated with CDG-2. Measuring their radial velocities would immediately test whether they share a common dynamical signature consistent with a bound dark-matter halo, or whether they are chance alignments within the Perseus intracluster light. Even a handful of precise velocity measurements could distinguish between a massive, dark-dominated system and a low-mass, tidally shredded remnant whose apparent darkness is an illusion of projection.
Deeper imaging will also matter. Longer Hubble or Euclid exposures, or future observations with the Nancy Grace Roman Space Telescope, could refine the surface-brightness profile of CDG-2 and search for additional globular clusters at larger radii. A rising globular-cluster count with distance from the center would support the picture of a substantial halo, while a sharp cutoff might hint at tidal truncation by the Perseus Cluster environment. Parallel searches using the same methodology in other rich clusters could reveal whether CDG-2 is unique or simply the first well-characterized member of a broader class of dark galaxies.
If CDG-2 is confirmed as a galaxy with an enormous dark-matter halo and almost no stars, it will sharpen the contrast with dark-matter-deficient systems and force theorists to explain how both extremes can emerge within the same cosmological framework. If, on the other hand, follow-up work shows that its dark character has been overstated, the case will still have served as a stringent test of how far current techniques can push into the low-surface-brightness frontier. Either way, CDG-2 underscores how much of the universe’s structure may still be hiding just below the limits of what our telescopes can reliably see.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
