Cosmic-web filaments — the elongated overdense bridges connecting galaxy clusters — were originally treated in ΛCDM as approximately symmetric structures whose linear mass density and galaxy content depend on distance from the spine and from the nearest cluster, but not on which end of the filament one approaches. Detailed photometric and spectroscopic mapping by Cautun et al. (2014), Bonnaire et al. (2020), and Gouin et al. (2023) now reveal a different pattern: many filaments exhibit a clear axial density gradient, with one end systematically denser — more thermalized gas, higher galaxy count, higher metallicity, hotter X-ray emission — than the other, and the gradient direction is not determined simply by the gravitational pull of the more massive endpoint cluster. Tudorache et al. (2025) extended the picture by detecting coherent bulk angular momentum along the filament axis — implying directional momentum at amplitudes 10–20× larger than ΛCDM tidal-torque theory predicts.
ΛCDM struggles to produce this asymmetry from its standard ingredients. Gaussian initial conditions plus tidal-shear fragmentation (Bond et al. 1996; Pogosyan et al. 2009) give symmetric or quasi-symmetric filaments; the Zel'dovich approximation predicts pancake-then-filament collapse with density profiles set by the local quadrupole of the tidal field, not by any preferred direction along the spine. The proposed remedies — late-time accretion biased toward the more massive endpoint, or ad-hoc "feeding rate" asymmetries from AGN feedback — neither predict the observed axial bulk rotation nor the observed metallicity gradient (which requires earlier, hotter processing on the denser end than on the sparser end). The combination of (a) axial density asymmetry, (b) coherent bulk rotation, and (c) metallicity gradient all pointing the same direction signals that filaments encode a directional formation history that pre-dates and survives subsequent infall.