Ever since Edwin Hubble discovered the fleeing of galaxies, the law of expansion of the Universe — Hubble's Law — has been a cornerstone of cosmology. Yet today, two main methods of measuring the Hubble constant diverge like the paths of ancient continents: the cosmic microwave background recorded by the Planck satellite gives about 67 km/s/Mpc, while calibration via supernovae by Adam Riess and the SH0ES team stubbornly points to 73. It’s not just a discrepancy — it’s a crack in the foundation, gaping at five standard deviations. Cosmologists call it the Hubble tension, and resolving it demands new physics.

One peacemaker is early dark energy (Early Dark Energy, EDE) — a short-lived component that, at the dawn of time, boosted expansion, shrinking the acoustic horizon and thereby shifting Planck’s predictions.

Meanwhile, new data on baryon acoustic oscillations from the DESI DR2 survey hint that today’s dark energy may also be inconstant: its equation of state w has varied smoothly with epoch. To stitch these threads together, physicists have turned to the deepest well — string theory, specifically its swampland hypothesis. According to the Distance Conjecture, at large scalar field displacements, the masses of particles, including dark matter, drop exponentially. Thus a natural link is born: the early dark energy field, as it evolves, can alter the mass of dark matter — and the latter invisibly responds, like the subterranean roots of a mangrove forest intertwining islands in a quaking bog.

In a new study, an international team performed a complex mapping of this theoretical landscape. Using conditional normalizing flows — an elegant machine learning method that instantaneously constructs posterior distributions in 14-dimensional space — the authors tested two popular EDE models: an axion-like one, where the field freezes at the bottom of the potential, and a model with an anti-de Sitter potential (AdS-EDE), where it keeps moving even in the late Universe. Both incorporated the exponential link between dark matter mass and the EDE field — exactly what the swampland hypothesis dictates. The data included all key modern surveys: Planck, DESI DR2 BAO, PantheonPlus, and SH0ES.

The results revealed a curious asymmetry. In the axion-like model, the coupling strength c is practically zero — the data see no reason for a link. But in AdS-EDE, the mean c is positive, and the fraction of coupled dark matter f* is above zero at the one-sigma level. The reason is that the continuing motion of the field allows a better sensing of this connection. In both models, the Hubble constant soared to ~72 km/s/Mpc, approaching the SH0ES value, but most intriguingly: the presence of coupling slightly lowered H0 while simultaneously strengthening the preference for evolving dark energy. In other words, the root from the mire stirs not only the surface water but also the deep currents that shape the late-time dynamics of the cosmos.

These findings echo the pioneering work of Saul Perlmutter and others, who proved accelerated expansion. Now, the details of the EDE potential, not just its presence, may prove decisive. The link between dark matter and energy, hinted at by fundamental principles of quantum gravity, ceases to be mere speculation — it becomes a testable connection that can unite two of the greatest mysteries. Ahead lie data from the Euclid telescope, the LSST surveys, and polarization maps from CMB-S4, which will either strengthen this bridge or point to new islands in the bog of theories. One thing is clear: the cosmological landscape, like the swamp, will surprise us more than once with hidden channels linking the most distant with the most profound. And although gigabytes of data still need sifting, it’s already apparent: the microscopic geometry of extra dimensions can cast shadows across the entire firmament.