Dark energy drives the accelerating expansion of the universe, and recent DESI results suggest it may change over time. Our independent research asks whether that change is tied to the growth of cosmic structure, and whether dark energy is spread evenly or gathers in particular places.
Before every test we write down what we expect and save it with a date and time. Then we run the test against public data (DESI, Planck, Pantheon+ supernovae and SDSS BOSS voids) and report the outcome, hits and misses alike.
If dark energy were denser in our own cosmic neighbourhood, local expansion would run faster and could explain the "Hubble tension", the disagreement between two measurements of the universe's expansion rate. We tested this against 1,657 supernovae, 77 of them calibrated by Cepheid stars, plus DESI and Planck data.
The early universe left a natural sound scale of about 100 Mpc/h. We predicted that dark energy might gather most in voids of that size, like a note resonating in a room of the right size. Stacking the Planck CMB map on 1,074 BOSS voids found only a faint signal, under 2 sigma. The size and edge patterns lean the predicted way; the wall patterns do not. Larger DESI void catalogues will settle it.
Fixed before the final DESI results and the Euclid mission's data. If they fail, we will say so here.
| Forecast | Value |
|---|---|
| Equation of state today | w = −0.94 (−0.97 to −0.92) |
| Redshift where w crosses −1 | z = 0.77 and z = 0.71 (two models); broader family z = 0.43–1.38 |
| w at redshift 2 | about −1.35 to −1.49 |
| CMB signal behind supervoids | peaks for voids of 80–130 Mpc/h, smaller peak near 50 Mpc/h |
What would prove us wrong: today's w measured below −1; no crossing of −1 between redshift 0.6 and 0.8; w at redshift 2 above about −1.2; or a supervoid signal that simply grows with void size.
Every forecast on this page comes out of models we built and tested step by step. Here are four of those steps, from early toy models to tests against real sky data. Misses are shown alongside hits.
Once bubbles of a new vacuum start to form (shaded band), dark energy weakens over time: w climbs from about −0.8 toward zero. That is the same direction recent DESI results hint at.
Regions with opposite character, scattered over a sphere, meet along seams. The seams close up over time; with more regions they linger far longer before vanishing.
Real data (1,657 supernovae, DESI, Planck) hold any extra dark energy in our own neighbourhood to 2–3%, far short of the 10–20% a Hubble-tension fix would need.
Behind 1,074 distant voids, the faint cold signal is strongest in the predicted 80–130 Mpc/h band, but no bin passes 2 sigma yet. DESI void catalogues will decide.
