Editorial NexoScience
In 1917, Albert Einstein added a term to his equations of general relativity to keep the universe still. He called it the cosmological constant, and he later treated it as a misstep. Eight decades later that same term came back as the simplest description of dark energy, the mysterious component pushing galaxies apart at an accelerating rate. Now the most precise survey of cosmic expansion ever assembled is hinting that the constant was never constant at all.
The data point toward dark energy that evolves. The Dark Energy Spectroscopic Instrument, mounted on the Mayall telescope at Kitt Peak in Arizona, is building a three-dimensional map of the cosmos by measuring the spectra of more than thirty million galaxies and quasars. Its second data release, analyzed alongside supernova catalogs and cosmic microwave background measurements, keeps landing on the same uncomfortable result. The equation-of-state parameter of dark energy, the number that describes how its pressure relates to its density, does not sit at the fixed value of minus one that a true constant demands. It appears to move across that line.
Crossing that line has a name, and a theoretical price. Physicists call a dark energy that crosses the minus one boundary a quintom, a hybrid of the two older families of models, quintessence above the line and phantom below it. A review led by Yifu Cai and colleagues walks through why this is so awkward. There is a no-go theorem in cosmology stating that no single ordinary scalar field and no single perfect fluid can smoothly cross that boundary. At the crossing point the effective sound speed of the fluid blows up and the perturbations become unstable. To describe what the data seem to show, theorists have to add something: a second field, higher-derivative terms in the Lagrangian, a modification of gravity itself through frameworks like f(R), f(T) or f(Q), or a direct energy exchange between dark energy and dark matter. Every one of those routes carries a cost, from ghost instabilities to extra degrees of freedom, and background-level observations alone cannot tell them apart.
The numbers are strong, and they are not yet decisive. Combined with the cosmic microwave background, the DESI DR2 baryon acoustic oscillation measurements prefer evolving dark energy over the standard model at about 3.1 sigma. Add supernovae and the preference climbs: 2.8 sigma with Pantheon+, 3.8 sigma with Union3, and up to 4.2 sigma with DESY5. Reconstructions that make no assumption about the shape of the curve reach 4.3 sigma in the best combination. Those are serious numbers, and the first DESI release pointed the same way at up to 3.9 sigma. But particle physics has long used 5 sigma as its discovery threshold, the bar the Higgs boson had to clear. Everything here sits below it. This is a persistent preference in the data, reproduced across different parameterizations and different reconstruction methods, and that consistency is what makes it interesting. It is not a confirmed detection.
The standard model already had cracks worth explaining. The cosmological constant carries a famously ugly problem: quantum field theory predicts a vacuum energy density more than 120 orders of magnitude larger than what is observed. There is also the coincidence problem, the oddity that dark energy and matter happen to have comparable densities right now, in our epoch. Add the Hubble tension, the sigma-8 tension and various large-angle anomalies in the microwave background, and the case for looking beyond a simple constant was building well before DESI.
The same mathematics could reach back to the beginning of time. The review notes a curious overlap. The kind of equation-of-state crossing that DESI seems to favor today is the same behavior required for bounce cosmologies, models in which the universe contracts to a finite minimum size and re-expands instead of emerging from an initial singularity. The tool built to explain cosmic acceleration turns out to be the tool that lets theorists write a universe without a Big Bang singularity, extendable to cyclic and emergent-universe scenarios. A dynamical dark-energy field could also couple weakly to ordinary matter, offering a route to explain the excess of matter over antimatter, or to photons through a Chern-Simons term that would slightly rotate their polarization on the way here.
The verdict depends on instruments that are already running. Later DESI releases will add statistics. Euclid and the Rubin Observatory will provide independent cross-checks with different systematics, which matters far more than another decimal place from the same telescope. Next-generation microwave background experiments including the Simons Observatory, CMB-S4 and AliCPT will tighten constraints further and could test that photon coupling directly. Until then, the honest summary is that the universe’s simplest ingredient is looking less simple, and cosmologists are watching closely rather than rewriting the textbooks.
Source: “The quintom theory of dark energy after DESI DR2”, by Cai Y, Ren X, Qiu T, Li M and Zhang X, published in National Science Review (2026). DOI: 10.1093/nsr/nwag115.
