The Hubble constant is supposed to be the same in every direction, and X-ray galaxy clusters say it is not. Migkas et al. (2020, 2021; arXiv:2103.13904) used scaling relations between cluster X-ray luminosity and temperature, calibrated independently of cosmology, to map the inferred H0 across the sky using more than 300 clusters. The result is a dipole-like anisotropy: H0 varies by about 9 percent between the fastest and slowest sky regions, with a roughly 4 sigma significance against isotropy when independent samples and relations are combined. The anisotropy direction is consistent across different scaling relations and cluster samples, and aligns approximately with directions previously flagged by supernova and bulk-flow analyses.
Within ΛCDM the expansion rate is isotropic by construction, so the signal must be repackaged as either an enormous coherent bulk flow, of order 900 km/s extending beyond 500 Mpc, far larger and deeper than the model's velocity field allows, or as unidentified systematics in the X-ray analysis. The bulk-flow escape is itself a falsification in disguise, since ΛCDM peculiar velocities should average away on those scales; CosmicFlows data independently show excess coherent flows in the same general direction, compounding rather than relieving the problem. The systematics defense weakens with each cross-check: the anisotropy survives changes of temperature calibration, sample selection, and relation choice.
The standing is provocative and underexamined: cluster scaling relations are an unconventional cosmological probe, and the community has been slow to confront a 4 sigma isotropy violation that, if real, invalidates the foundational symmetry assumption of every FLRW-based analysis. eROSITA's all-sky cluster survey is the designed referee, with the statistics to confirm or refute directional H0 variation at high significance.