Gravitational lenses are nature's own Hubble experiment, and their verdict started a war over a degeneracy. When a quasar sits behind a massive galaxy, its multiple images arrive with delays set by the lens geometry and the expansion rate: measure the delays, model the lens, and H0 falls out with no distance ladder, no CMB, no supernovae. The H0LiCOW program did exactly that with six lensed quasars and reported H0 = 73.3 (+1.7/-1.8) km/s/Mpc (Wong et al. 2020), agreeing with the local ladder at 2 percent precision and standing 3 sigma from Planck: an independent vote for the high value, from physics with completely different systematics. The result sharpened the Hubble tension materially, until the mass-sheet degeneracy reasserted itself: the TDCOSMO reanalysis (Birrer et al. 2020) showed that relaxing the lens-profile assumptions H0LiCOW had adopted lets the same data support H0 anywhere from the mid-60s to the mid-70s, with the answer controlled by stellar-kinematic priors rather than the delays themselves.
The episode is now a microcosm of the wider crisis: a ladder-independent method lands high when modeled with astrophysically motivated profiles, then relaxes toward ambiguity when its absolute mass scale is left free, and the choice between those treatments imports exactly the question under adjudication. Meanwhile the lenses themselves sit in varied line-of-sight environments whose convergence corrections are modeled statistically, and the method's future, time delays from lensed supernovae, hundreds of new systems from Rubin and Euclid, spatially resolved kinematics from JWST, will either converge the degeneracy or expose an environmental structure the single-number framework cannot hold.
The standing is a precision method mid-arbitration: the high central values persist in the anchored analyses, the degeneracy-freed envelope spans the tension, and the decisive kinematic data are arriving now.