On 11 June 2026, the journal Monthly Notices of the Royal Astronomical Society reportedly published two papers about the same 1,700-odd exploding stars and reached flatly opposite verdicts.
One, led by the Tata Institute of Fundamental Research with Oxford's Subir Sarkar, is titled "Pantheon+ supernovae corrected for progenitor age indicate the universe is decelerating."
The other, led by Southampton's Phil Wiseman with Oxford's Maria Vincenzi and Nobel laureate Adam Riess, is reportedly titled "Still accelerating: type Ia supernova cosmology is robust to host galaxy age evolution."
Same issue. Same dataset. Oxford names on both sides. A rare public split, argued in print, over whether the central discovery of modern cosmology is real.
Same issue. Same dataset. Oxford names on both sides — a rare public split over whether the central discovery of modern cosmology is real.
The coverage has mostly framed this as a debunking and its rebuttal: Sarkar's team says dark energy is an "illusion," and the establishment says, no, the evidence still holds. That framing gets the physics backwards.
Because Sarkar's letter is not one claim. It is two — and they are not equally strong. Separating them is the whole game.
The first claim is about brightness. Type Ia supernovae are used as standard candles because they are thought to emit a fixed amount of light, so their apparent dimness tells you their distance. But a South Korean group at Yonsei has been building evidence that the light depends on the age of the star that exploded: younger progenitors are fainter at fixed stretch and color.
Sah, Rameez and Sarkar applied a redshift-dependent correction for that age effect — reportedly roughly 0.03 magnitudes per billion years of progenitor-age drift — to the Pantheon+ catalogue.
The result, in the paper's own technical language, is that the correction "shifts the monopole component of q0 to positive values (i.e. deceleration)." The monopole is the direction-averaged expansion. Push it positive and the universe, on average, is slowing down. No acceleration, no need for dark energy.
That is the headline. It is also the contested, potentially fixable part.
Wiseman and Vincenzi's rebuttal targets exactly this. Their argument is that the age effect is real but already absorbed. As Vincenzi puts it, "we have long known that the brightness of Type Ia supernovae depends on the age of the stars that produce them" — and modern analyses correct for it indirectly, through the mass of the host galaxy.
Sarkar's own paper concedes the fight is live. In a footnote, it notes that Wiseman's team argues host-mass standardization "removes any correlation between the Hubble residuals and the progenitor age," and adds, plainly: "The debate continues."
So the deceleration result hangs on a specific, technical dispute about whether a stellar-age correction is double-counting something already captured by host-galaxy mass. That question can be adjudicated with better data and better bookkeeping. It is a normal scientific argument. Dark energy could survive it.
Now the second claim. This is the one being under-reported, and it is the one that actually threatens the paradigm.
Sarkar's team also asked whether the inferred acceleration looks the same in every direction. It does not. In their tomographic analysis, the deceleration parameter q0 is, reportedly per their earlier Sah et al. 2025 analysis, "mainly a dipole which dies out by ∼0.1" in redshift — meaning the effect points in a preferred direction and fades with distance, vanishing beyond a few hundred megaparsecs.
That direction is not random. The dipole aligns approximately with the local bulk flow — the large-scale streaming motion our galaxy and its neighbors are caught in — and it is most pronounced when measured in the frame of the Local Group. It points, roughly, the way we are already moving, marked by the hotspot in the cosmic microwave background.
Here is why that is fatal in a way the brightness dispute is not. Dark energy, in the standard picture, is a property of the quantum vacuum. It is the same everywhere and in every direction. It has no preferred axis. An effect that points somewhere cannot be it.
And crucially, this claim does not depend on the age correction at all. In Sarkar's words: "This is unaffected by the correction to the supernova brightness — so rejects dark energy independently of whether the correction is applied or not. The correction turns the isotropic component into a deceleration — which again rules out dark energy."
Read that carefully. He is describing two separate kill shots. The age correction handles the monopole — the average. The dipole handles the direction. And the dipole survives even if you throw the age correction out entirely.
This is the load-bearing part of the argument, and it is not really about supernova brightness. It is about whether the universe is lopsided.
Sarkar's broader case leans on this. The standard analyses of the CMB and of baryon acoustic oscillations all run through the FLRW framework, which assumes the universe is, on average, the same in all directions. His group claims to have evidence that "falsifies the FLRW framework" — that the universe is asymmetric.
They are not alone in seeing a preferred direction. The claimed anisotropy in the distribution of distant quasars and radio galaxies — the cosmic dipole anomaly — is, per the Nobel laureate Jim Peebles, about as well established as the Hubble tension, even if it has drawn a fraction of the attention.
The interpretation offered is that we are "tilted observers." Embedded in a bulk flow, our local motion produces an apparent acceleration along that flow — a general-relativistic artifact of where we sit, not a vacuum energy filling all of space.
You can disagree with that interpretation. Many cosmologists do. Joshua Frieman at SLAC notes the robust, independent evidence for acceleration from large-scale structure and the CMB, and argues that questioning one line of supernova evidence is not "sufficient to overthrow the paradigm."
But notice what the establishment rebuttal actually addressed. Wiseman and Vincenzi showed that the supernova brightness corrections are "already understood and accounted for." Riess framed it well: "Extraordinary claims require especially careful testing."
That testing landed squarely on the monopole — the brightness question, the deceleration result. It is the answer to the claim that the field could most cleanly rebut.
The dipole is a different animal. A brightness recalibration does not make a directional signal isotropic. If the acceleration genuinely points along our bulk flow and dies out with distance, no amount of host-mass bookkeeping repairs that, because it is a claim about geometry, not luminosity.
Which reframes what the coming decade of data has to settle. The Rubin Observatory's Legacy Survey of Space and Time will measure hundreds of thousands of supernovae — and both camps are already pointing to it as the arbiter.
More supernovae will sharpen the monopole. They will pin down whether the average expansion is accelerating or slowing, and whether the age correction is real or redundant. That is the fight everyone is watching.
But the deeper test is isotropy. The question Rubin most needs to answer is not "how bright, on average" but "is the sky the same in every direction, or does the acceleration have an axis?"
If it has an axis — if the universe is genuinely lopsided at the scale these supernovae probe — then dark energy as a uniform vacuum energy is in real trouble, no matter how the brightness dispute resolves.
And it is worth remembering that the paradigm being defended is not as settled as the debunking framing implies. The largest survey of physicists ever conducted reportedly found that ΛCDM, the standard model built around dark energy and cold dark matter, failed to win majority support.
That does not make Sarkar right. It means there is no consensus here to overthrow — only a working framework with cracks that specialists argue over in public.
The two papers in that June issue are not a claim and its refutation. They are two teams answering two different questions and mistaking them for one.
Whether the universe is lopsided is the one that will decide it. And that is not a question about how bright a supernova is.