ΛCDM's small-scale bookkeeping predicts far more satellite galaxies than the sky contains. High-resolution N-body simulations of Milky-Way-mass halos produce hundreds of dark matter subhalos massive enough, in principle, to host dwarf galaxies, with classic counts predicting of order 500 against the roughly 60 satellites known around the Milky Way; in the dwarf-capable mass range commonly quoted the ledger reads about 50 expected luminous satellites against roughly 20 observed bright ones. Deep surveys (SDSS, DES, Pan-STARRS) have added ultra-faint members and narrowed the raw gap, while simultaneously sharpening its anisotropy: the satellites that do exist arrange themselves in thin co-rotating planes rather than the isotropic swarm subhalos predict.
The standard resolution makes most subhalos dark: reionization heating and supernova feedback prevent star formation in small halos, so the missing satellites are present but invisible. This rescue is quantitatively flexible and qualitatively costly. It converts the prediction into an unfalsifiable population, requires the suppression threshold tuned to track the observed luminosity function, collides with the too-big-to-fail problem (the most massive predicted subhalos are too dense to match any observed satellite, and those should not be dark), and does nothing for the planes. After two decades the missing-satellite account stands only by coupling to additional patches, each carrying its own tension.
The standing has evolved from counting crisis to structure crisis: the number can be argued, the phase-space arrangement cannot. Rubin-LSST will complete the ultra-faint census and fix the satellite luminosity function to the survey horizon, closing the bookkeeping question while leaving the deeper one: why the survivors orbit in organized planes that random subhalo accretion cannot supply.