Japan-led supernova reanalysis challenges dark energy certainty

The accelerating universe may not be accelerating at all

The standard model of cosmology rests on a single, quietly radical assumption: something invisible is pushing the cosmos apart faster and faster. That something — dark energy — accounts for roughly 68 percent of the universe’s total energy budget and has been treated as settled science since 1998, when two independent teams spotted distant Type Ia supernovae dimmer than expected.

A new reanalysis is now challenging that consensus, and it is doing so with more rigor than most controversies in this field attract.

Professor Subir Sarkar and his team examined the Pantheon+ catalog, which contains brightness measurements from 1,550 Type Ia supernovae and 1,701 individual light-curve records. Their argument is deceptively simple: if the age of the white dwarf progenitor star before it explodes affects how bright the supernova truly is, then existing corrections may have systematically biased distance estimates — and with them, the entire case for cosmic acceleration.

The logic traces back to earlier observational work showing that supernovae in galaxies dominated by young, actively star-forming stars tend to appear dimmer than those in older, quiescent populations. Previous researchers attributed this dimming primarily to metallicity — the concentration of elements heavier than helium in the progenitor system — because younger galaxies tend to be more metal-poor. But Sarkar’s group asked what happens if you correct for progenitor stellar age explicitly, using redshift-dependent averages drawn from those same earlier studies.

What they found turns the standard picture inside out. After applying the age correction, the inferred cosmic expansion history shifts from acceleration to deceleration. The universe is still expanding, but its rate is slowing — the opposite of what dark energy predicts.

That result alone would be shocking. But the data also preserves something unexpected: a directional signal. Even after correction, supernovae in certain sky regions still appear to show signs of accelerated expansion. The effect is unchanged in strength from the uncorrected analysis, suggesting it is not an artifact of the correction itself but rather a real feature embedded in the observations.

Sarkar’s team interprets this residual signal as evidence of a bulk flow — the Milky Way, our galaxy group, and the surrounding supercluster moving together through space at a significant velocity relative to the Hubble flow. If our vantage point is in motion, objects in the direction of that motion will appear slightly blueshifted relative to those behind us, creating an asymmetric brightness pattern that mimics the signature of cosmic acceleration without requiring any new physics. The analysis estimates a peculiar velocity of roughly 600 to 700 kilometers per second in the direction of the constellation Centaurus, consistent with other large-scale flow measurements but large enough to shift distance inferences at the precision level required by dark energy studies.

This is not the first time someone has pointed to a dipole anomaly in supernova data. In 2023, a similar analysis by the same group used a smaller sample and attracted immediate skepticism. What distinguishes the current work is the sample size and the explicit separation of isotropic expansion from directional effects — something previous analyses largely glossed over. By fitting the expansion rate independently in different sky directions, the researchers demonstrated that the dipole persists even when the analysis accounts for statistical noise, which had been the primary counterargument to earlier claims.

The pushback is already here

Not everyone is convinced. A team at the University of Southampton led by Phil Wiseman tested Sarkar’s correction using standard luminosity calibration methods that already account for host galaxy stellar mass, star formation rate, and other environmental factors baked into the Pantheon+ analysis pipeline.

Their conclusion, published as a companion study, was blunt: once those conventional corrections are applied, the apparent relationship between progenitor age and supernova brightness disappears. In other words, the effect Sarkar’s group identifies may already be absorbed into the standard analysis — and reapplying it double-counts a factor that is no longer present in the residuals.

Wiseman’s team also raised a more fundamental objection. The average stellar age of a host galaxy is not the same as the age of the specific white dwarf that explodes. Correlating one with the other introduces an uncertainty that could easily swamp the signal the researchers claim to have found. A galaxy may be old on average while harboring a recently formed star, and the supernova’s progenitor white dwarf may have formed through a channel entirely unrelated to the galaxy’s bulk stellar population.

The debate is real and unresolved. Sarkar acknowledges it in the paper, noting that the relationship between progenitor age and supernova brightness remains contested. But acknowledging a dispute is not the same as resolving it, and the question of who is right will likely dominate observational cosmology for the next several years.

Why this matters beyond academia

The stakes extend well beyond a technical disagreement about light curves. The entire dark energy research enterprise — from ground-based spectrographs like DESI and the Antarctic Dark Energy Explorer to the Euclid space mission and proposals for the proposed Wide Field Survey Telescope — is designed to measure the equation of state of dark energy with ever-greater precision. If the acceleration signal is partially or wholly an artifact of progenitor system properties, much of that investment is pursuing a mirage that could dissipate under better-calibrated data.

It also matters for the Lambda-CDM model itself. That framework, which has survived two decades of increasingly precise tests from the cosmic microwave background, baryon acoustic oscillations, and large-scale structure surveys, would need revision. Not abandonment — multiple independent probes still converge on a universe dominated by dark energy — but a genuine modification. Dark energy might not be a cosmological constant. It might not exist at all as a separate component. It might instead reflect a misunderstanding of how we calibrate our primary distance ladder.

There are second-order implications as well. If cosmic acceleration is partially or wholly a kinematic artifact, it reshapes how we interpret the integrated Sachs-Wolfe effect, the growth rate of cosmic structures, and the tension between early- and late-universe measurements of the Hubble constant. It also opens the door to alternative gravity theories that were previously constrained to the margins, since a reduced need for dark energy loosens the tight bonds between expansion history and structure formation.

The data coming next could settle it

The Vera C. Rubin Observatory’s Legacy Survey of Space and Time, which begins full operations this decade, will detect hundreds of thousands of Type Ia supernovae — roughly a thousand times the Pantheon+ sample. That volume of data will allow progenitor-age corrections to be tested directly rather than inferred indirectly from host galaxy properties. For the first time, researchers will have enough supernovae with well-characterized progenitor systems to distinguish whether brightness variations track stellar population age or something else entirely.

If Sarkar’s correction holds up at that scale, the standard model gets rewritten. If it doesn’t, the dark energy paradigm gains a new layer of robustness that makes future challenges far harder to mount.

Neither outcome is guaranteed. But for the first time in years, the accelerating universe is no longer an assumption — it is a hypothesis under active scrutiny, and the scrutiny is coming from a direction that the cosmology community can no longer ignore. The implication is clear: the biggest unknown in physics may not be what dark energy is, but whether it is there at all.