A cold, black mirror — that's what the sea of dark energy, the invisible substance driving the expansion of the universe, looked like for billions of years. But the DESI spectrograph has caught a faint ripple. Frozen sound waves — baryon acoustic oscillations — hinted that cosmic distances are off-kilter. As if an invisible tide weakens with each epoch, and the redshift of light from distant galaxies grows slightly differently than Hubble's law predicts. The accelerated expansion, discovered by Adam Riess, Saul Perlmutter, and Brian Schmidt from supernova brightness, may not fit the simplest scenario.

The usual explanation is time evolution. But time has a rival: space. And its curvature can play a nasty trick. If our Galaxy floats inside a giant bubble with reduced density of dark energy, an observer would be inside a cosmic lens. The scenario isn't pulled out of thin air: in quantum field theory, the vacuum can be metastable. A phase transition, like a bubble in boiling water, creates a region of new phase expanding at nearly the speed of light. The authors built exactly such a model: a spherical bubble inside, where the density of dark energy is 10% lower than outside, nucleated at a redshift of 1,4 — just 4,5 billion years after the Big Bang.

But the universe has a relentless timekeeper: the cosmic microwave background, the afterglow of the Big Bang. Its near-perfect homogeneity was set in the first thousand years. A bubble that formed later must distort that picture. If we are not at the geometric center (the odds are minuscule), other points inside the bubble see their own dipole of the cosmic microwave background. Electrons in galaxy clusters scatter this dipole, producing the kinetic Sunyaev–Zel'dovich effect (kSZ). Calculations showed its amplitude on angular scale ℓ=3000 should be hundreds of μK². But observations give only 2,9 μK² — a gap of two orders of magnitude.

The model is shattered. The sea of dark energy turned out to be stormy, and bubbles too conspicuous. But a negative result maps the minefield. Now it's clear: inhomogeneous dark energy leaves indelible imprints in the CMB and large-scale structure. Future surveys like Euclid will either find them or close the last loopholes. Perhaps bubble walls are thicker, motion subsonic; perhaps we are in an ensemble of tiny bubbles. The question of the constancy of the cosmological constant remains open.