Astronomers combined three large surveys to map how mass is distributed from the centers of galaxies out to their dark matter halos. They find that the biggest halos have noticeably less dark matter inside their inner regions than popular computer simulations predict. This difference shows up when comparing dynamical measurements at different radii and persists after accounting for the visible matter.

The team used stellar motions measured by the MaNGA survey to probe the inner galaxy, neutral hydrogen (H I) 21-cm measurements from the ALFALFA survey to probe larger radii, and independently calibrated halo masses for galaxy groups from SDSS to fix the total halo scale. The MaNGA work uses a method called Jeans Anisotropic Modelling to estimate the enclosed dynamical mass at the three‑dimensional half‑light radius. The ALFALFA data give a dynamical mass at the H I radius. Because the H I measurements alone cannot reliably give the total halo mass for massive systems, the independent group-based halo masses were essential.

When these anchors are combined, the observational mass profiles agree fairly well with the IllustrisTNG and EAGLE simulations for low‑mass halos. For massive halos, however, the observations show systematically lower dynamical masses at the H I radius, lower dark‑matter masses in the inner regions, and lower central dark‑matter fractions. Two independent observational estimates of the central dark‑matter fraction both show the same trend. The offset in the most massive bins is large compared with the population spread—about a four‑sigma difference in units of that scatter.

After subtracting the contributions from stars and gas, the inferred dark‑matter profiles still look broadly like the common Navarro–Frenk–White (NFW) shape used to describe dark‑matter halos. The main difference is that the effective concentration—the degree to which dark matter is packed toward the center—is lower in massive observed halos than predicted by the simulations. The authors suggest this points to stronger long‑term baryonic ‘‘halo heating’’ in real massive systems than current simulations reproduce. Such heating can come from processes linked to stars, active galactic nuclei, or interactions with satellites and mergers, which can move dark matter outward.