IceCube has measured an astrophysical neutrino flux from TeV to PeV energies for over a decade, and the sky map refuses to match the catalogs. The diffuse flux is consistent with isotropy, implying a dominant extragalactic origin, yet the source populations expected to produce it keep failing association tests: gamma-ray-bright blazars, the presumptive accelerators, are constrained by stacking analyses to contribute at most 10 to 30 percent of the flux; the identified candidate sources are bafflingly heterogeneous (the blazar TXS 0506+056, the Seyfert NGC 1068, the Galactic plane detected at 4.5 sigma in 2023); and NGC 1068's neutrino emission exceeds its gamma-ray output, requiring a calorimetric, gamma-obscured core unlike the transparent jets models favored. The bulk of the neutrino sky remains unattributed: whatever produces most of the flux is either numerous and individually faint, or hidden at gamma-ray energies, or both.
For ΛCDM this is an astrophysical inventory problem with the familiar shape: the catalogued populations underproduce, the required hidden population is postulated after the fact, and each association (a tidal disruption event here, an obscured AGN there) adds heterogeneity rather than resolution. The neutrino-gamma tension is structural: the same hadronic processes that make neutrinos make gamma rays, so the missing gamma-ray counterparts demand sources opaque to their own photons, a class the standard census did not anticipate in the required numbers.
The standing is an open attribution crisis entering its precision era: IceCube-Gen2 and KM3NeT will resolve the flux into sources or bound their faintness, and the hidden-core population the data demand is becoming a definite, testable inventory requirement.