Dark flow was the claim that everything is drifting somewhere, and its career traces how ΛCDM handles motion it cannot source. Kashlinsky and collaborators (2008 onward) stacked the kinetic Sunyaev-Zeldovich imprints of hundreds of galaxy clusters in WMAP data and reported a coherent bulk motion of 600 to 1000 km/s extending beyond 300 Mpc toward (l, b) near (290, 30), a flow with no visible attractor inside the volume: matter beyond the observable horizon, they argued, was pulling. The claim was contested from birth, Planck's 2013 analysis with better data and the same method found no significant dark flow, and the original detection is now widely regarded as a systematics casualty, but the question it raised did not retire with it.
The surviving form is the bulk-flow excess: direct peculiar-velocity surveys, culminating in CosmicFlows-4 analyses (Watkins et al. 2023), measure coherent flows of about 400 km/s on scales of 200 to 250 Mpc, two to three times the ΛCDM expectation for those depths, with the discrepancy's significance in the 3-to-5 sigma range depending on estimator. The model's velocity field is fully determined by its matter distribution and initial power spectrum: flows this fast, this deep, require either more large-scale power than the CMB-normalized spectrum contains or sources of motion the inventory lacks. Whether the residual flow points along the original dark-flow direction, toward the Shapley concentration and beyond, remains an active question entangled with the CMB dipole's own anomalies.
The standing is a disputed headline with a confirmed residue: the original dark flow is likely dead, the excess bulk flow that replaced it is alive at increasing significance, and the model is short of attractors either way.