Dark energy may not be constant, and that would rewrite the universe's story
For decades physics assumed dark energy, the force pulling the cosmos apart, is a fixed constant. New data from the DESI survey keeps finding a persistent hint that it changes over time. If it holds up it overturns the standard model of cosmology. The catch: it is a hint, not a discovery, and honesty about the difference is the whole story.
Dark energy makes up about 70% of the universe and drives its accelerating expansion, and the standard model of cosmology assumes it is a cosmological "constant" that never changes. On 30 July 2026, the DESI collaboration released a new analysis that adds an independent line of evidence that dark energy may instead be evolving, weakening over cosmic time. Combined with other data, that preference now sits at roughly 2.7 to 3.2 sigma, a persistent hint but well short of the 5-sigma threshold physicists require to call something a discovery. If it is real, it would be one of the biggest results in modern physics. If it is a statistical mirage, it will fade. Right now it is genuinely, honestly, unresolved.
There is a real chance that one of the deepest assumptions in physics is wrong, and we are watching the evidence build in real time. Dark energy, the mysterious thing accelerating the expansion of the universe, has always been treated as a fixed constant. A major cosmic survey keeps producing data that hints it might not be. The latest instalment landed on 30 July 2026, and it is exciting precisely because of how careful you have to be about it. Here is what was found, what it would mean, and why "hint" is the most important word in this story.
What is dark energy, and why does "constant" matter?
Dark energy is the placeholder name for whatever is causing the universe's expansion to speed up rather than slow down. It is not a minor detail: it accounts for roughly 70% of everything in the cosmos, outweighing all the matter, stars and galaxies combined. We do not know what it is.
The reigning model of cosmology, called Lambda-CDM, makes a specific assumption about it: that dark energy is a cosmological constant, an unchanging energy baked into empty space itself. In the physics shorthand, its "equation of state" (a number written w) equals exactly −1 and never budges, meaning its density stays the same as the universe expands. This assumption is a load-bearing wall of modern physics. If dark energy instead changes over time, that wall cracks, and much of what we think we know about the universe's past and future has to be rebuilt.
What did DESI actually find?
The Dark Energy Spectroscopic Instrument (DESI) is built to test exactly this. Its newest analysis, released on 30 July 2026, used a technique called the Lyman-alpha forest: the way light from more than 820,000 distant quasars is imprinted by the hydrogen gas it passes through on its way to us, letting astronomers map the universe's expansion as it was around 11 billion years ago.
On its own, that map is consistent with several possibilities. But when DESI's data are combined with two other pillars of cosmology, the leftover glow of the Big Bang (the cosmic microwave background) and exploding stars used as distance markers (supernovae), the numbers lean a particular way. A dark energy that evolves over time is preferred over a constant one at about 2.7 sigma using DESI and the microwave background, rising to 3.2 sigma when the supernovae are added, according to the collaboration's own paper. In the best-fit picture, dark energy was slightly stronger in the past and has been weakening.
Crucially, this is not a bolt from the blue. It is the latest in a building series: DESI first reported this hint in 2024, strengthened it with a bigger dataset in 2025, and has now added a technically independent probe. The signal has been stubbornly persistent, which is what makes it interesting.
So is this a discovery? No. And that matters enormously.
Here is where honesty separates science from headlines. This is not a discovery, and the DESI team is careful to say so. Physics has a demanding standard for claiming a new result is real: 5 sigma, meaning the odds of it being a fluke are about one in 3.5 million. This result is at 2.7 to 3.2 sigma. That is a meaningful hint, not proof, and results at that level do sometimes evaporate as more data arrives.
Two more caveats keep it honest, and they are the crux of the whole debate:
- The signal only appears in combination. DESI's expansion map alone does not cross these thresholds; the significance emerges only when it is stacked with the microwave background and supernovae. The evidence is a chorus, not a soloist.
- The supernovae matter, a lot. The exact significance depends on which catalogue of supernovae you use. In DESI's 2025 results, swapping one supernova sample for another moved the number from under 3 sigma to over 4. When your headline conclusion swings with the choice of dataset, caution is not optional.
There is already a healthy skeptical literature probing whether the effect is real physics or a subtle artefact of analysis choices. That is not a knock on DESI; it is science working as it should.
What would it mean if it is real?
If dark energy genuinely changes, the consequences are enormous, which is exactly why nobody should announce it prematurely.
It would break Lambda-CDM, the model underpinning nearly all of modern cosmology, and force a rewrite of how the universe evolved. It would also reshape predictions about the universe's ultimate fate. The standard picture has dark energy driving an eternal, accelerating expansion into cold emptiness. If instead dark energy is weakening, that far-future story is no longer certain, and more exotic endings come back onto the table. Those are conditional consequences, worth stating clearly precisely because the "if" is doing so much work.
What is DESI?
Worth knowing, because the instrument is part of why this is credible. DESI sits on the Mayall four-metre telescope at Kitt Peak in Arizona, run by a large international collaboration managed by the US Department of Energy's Berkeley Lab. It uses about 5,000 robotic fibre-optic "eyes" that reposition between exposures to capture the spectra of thousands of galaxies and quasars at once, on its way to mapping tens of millions of them. By measuring a subtle regularity in how matter is distributed (baryon acoustic oscillations), it charts how fast the universe expanded at different epochs, which is precisely the measurement that pins down dark energy's behaviour.
The evidence, building
| When | Data | Preference for evolving dark energy |
|---|---|---|
| 2024 (DR1) | First-year map, ~6 million redshifts | First hints |
| March 2025 (DR2) | ~14 million measurements | ~2.8 to 4.2 sigma (depending on supernova sample) |
| July 2026 | Lyman-alpha forest, 820,000+ quasars | 2.7 sigma (with CMB), 3.2 sigma (+ supernovae) |
The honest bottom line
Something interesting is happening at the foundations of cosmology, and it deserves neither a coronation nor a dismissal. DESI keeps finding, across independent methods, that a changing dark energy fits its data slightly better than a constant one, and the hint has held up rather than faded. That is genuinely exciting. It is also, by physics' own strict standard, not yet a discovery, it leans on combining datasets, and it is sensitive to choices that reasonable scientists argue about. The correct posture is the one the field itself is taking: pay very close attention, and wait for the data to cross the line or fall back. Over the next few years, as DESI and other surveys report more, one of two things will happen: a hint will harden into one of the great discoveries of the century, or it will quietly dissolve. Both are how science is supposed to work. For more from the frontier, see the Science section and our look at quantum computing's real breakthrough and the hype it doesn't justify.
For the first time, we can see whirlpools on the Sun's surface. They may explain why its atmosphere is so strangely hot
The world's largest solar telescope has taken the sharpest images ever of the Sun's surface and caught something long predicted but never seen there: tiny plasma whirlpools, only about 20 km across, the same wave-curl you get in ocean waves and clouds. They may be part of the answer to a decades-old mystery. Here is what was found, and what it does and does not prove.
JWST found a 'black hole star' in the early universe, and it might solve a real puzzle
The James Webb telescope spotted an object 660 million years after the Big Bang that looks like a giant star but is really a black hole wrapped in so much gas it disguises itself. If the interpretation holds, it helps explain one of astronomy's genuine mysteries: how monster black holes appeared so impossibly early. Here is the careful version.
A blood test can now help diagnose Alzheimer's. Here's what that really means
Diagnosing Alzheimer's used to require an expensive brain scan or a spinal tap. In 2026 a simple blood test crossed from the research lab into the clinic, sharply improving how accurately doctors can identify the disease. It is a genuine turning point, and also not a crystal ball. Here is what these tests do, and what they do not.