Dark Energy May Not Be Constant — And That Changes Everything

The standard model of cosmology has a problem. Again.

For thirty years, physicists have built their understanding of the universe on the Lambda cold dark matter model — ΛCDM — a framework that treats dark energy as a fixed, unchanging property of space itself. The cosmological constant Λ is supposed to be constant. Period. It was the simplest explanation, and by every measure that matters in physics, simplicity is a virtue. Occam’s razor has guided theoretical work since the mid-twentieth century, and ΛCDM passed that test with elegance. It required fewer assumptions than any competing model while explaining an extraordinary range of observations.

That assumption just got harder to justify.

The latest round of data from the Dark Energy Spectroscopic Instrument, or DESI, combined with observations from the cosmic microwave background and distant supernovae, suggests dark energy may not be constant at all. It may be evolving. If that finding holds up — and the statistical confidence levels are building, though not yet at the threshold that would constitute a discovery — it does more than tweak a parameter. It undermines the simplest version of the standard model that has guided cosmology since the late 1990s. The implications extend far beyond a single number. They cut into how we understand vacuum energy, quantum field theory, and the ultimate fate of everything that exists.

The Hubble tension was only the beginning

The crack started with a number. Or rather, two numbers that should have agreed and do not.

The Planck satellite, measuring the afterglow of the Big Bang, calculated the universe’s expansion rate at 67 kilometers per second per megaparsec. Adam Riess’ SH0ES team, measuring distances to nearby galaxies using Type Ia supernovae calibrated by Cepheid variable stars, found 73. Six kilometers per second apart. That gap has widened with better instruments instead of narrowing. Each generation of measurement has come with tighter error bars, and the conflict has only deepened.

Scientists called it the Hubble tension and treated it as an annoying discrepancy — something to be resolved as measurements improved. It did not improve away. It grew more persistent. And then it stopped looking like a measurement error. The probability that this is a fluke — that two independent methods, using completely different techniques and observing different epochs of the universe, simply converged on wrong answers by chance — is now estimated at roughly one in ten thousand. That is not noise. That is a signal.

The Hubble tension forced cosmologists to confront a possibility they had long dismissed: that the model itself was flawed, not just the measurements. An evolving dark energy is one of several proposed solutions, though not the only one. Some have suggested new relativistic particles in the early universe. Others have pointed to modifications of general relativity. But evolving dark energy has gained traction because it is the most straightforward adjustment — it requires changing only one assumption rather than rebuilding the entire theoretical framework.

DESI threw a wrench into the works

In March 2025, DESI researchers released findings that reframed the entire conversation. The instrument maps millions of galaxies and quasars to reconstruct cosmic expansion across billions of years, using baryon acoustic oscillations — the frozen imprint of sound waves from the early universe — as a standard ruler. Its data, when combined with other observations, indicated that dark energy might be changing over time rather than remaining fixed.

This matters because ΛCDM depends on Λ being constant. If dark energy evolves, the cosmological constant is not the full story. The simplest model is wrong. Not approximately wrong. Fundamentally incomplete. The cosmological constant was introduced by Einstein as a mathematical fudge factor, later abandoned, then resurrected in the 1990s as the leading explanation for cosmic acceleration. It fit the data beautifully. Until it didn’t.

Riess and fellow astronomer Alexie Leauthaud captured the mood in a paper published in August 2025: “Widening cracks are appearing in the Λ cold dark matter model.” They went further, noting that the standard model struggles to describe the full expansion history of the universe when confronted by all available data simultaneously. The language was careful — physicists rarely speak this bluntly — but the message was unmistakable. Something is wrong with the foundation.

Then came the Lyman-alpha forest complication

Science rarely moves in a straight line. In July 2026, DESI researchers released a new analysis using the Lyman-alpha forest — the pattern of absorption imprinted on ancient quasar light as it passes through hydrogen clouds across the cosmos. This technique probes a different regime than baryon acoustic oscillations, looking further back in time and sampling a different volume of the universe.

The result moved the expansion measurement closer to ΛCDM’s predictions. More precise, yes. But also more complicated. It does not resolve the tension. It muddies the picture further. The Lyman-alpha data, interpreted within the standard model, yields an expansion rate consistent with Planck. But when interpreted alongside the BAO results, the combined dataset still favors a time-varying dark energy equation of state. The tension persists, now split across multiple observational pathways rather than concentrated in a single discrepancy.

One dataset suggests dark energy is evolving. Another pulls back toward constancy. The truth is likely somewhere in between, or neither, and the data alone are not yet decisive enough to declare a winner. What is clear is that the era of trusting a single measurement to settle fundamental questions is over. Cosmology is entering a phase where competing datasets demand reconciliation, and no single instrument can claim authority.

Who wins and who loses

If dark energy is evolving, the implications cascade through every corner of cosmology. The theoretical consequences are immediate and far-reaching. The cosmological constant, currently understood as the energy density of the vacuum, would need to be replaced by a dynamical field — a quintessence field or something more exotic. This would require revisions to quantum field theory, which predicts a vacuum energy enormous compared to what we observe — the worst theoretical prediction in the history of physics. An evolving dark energy might finally provide a handle on that discrepancy, or it might deepen it.

The fate of the universe becomes an open question again. A constant dark energy drives perpetual acceleration — a cold, lonely expansion that tears galaxies apart over trillions of years. Evolving dark energy opens other possibilities. It could weaken, allowing gravity to reassert dominance and potentially leading to a contraction scenario. It could strengthen, accelerating expansion beyond current projections and producing a violent Big Rip. The standard narrative of a heat death is no longer guaranteed. The end of everything is no longer written in stone, or rather, in the cosmological constant.

The Nobel Prize winners themselves are sounding alarms. Riess, Perlmutter, and Brian Schmidt earned the 2011 Physics Nobel for discovering cosmic acceleration. Their work sits at the foundation of ΛCDM. Now Riess is openly questioning the model that grew from that discovery. When the architects of a theory question its core assumptions, it is worth listening. This is not contrarianism. It is the scientific method functioning as designed — curiosity overriding attachment.

On the losing side: physicists who built careers around refining the constant-dark-energy framework. Not because their work was wasted — it produced real knowledge, precise constraints, and decades of predictive success — but because the destination keeps moving. The simplest explanation was elegant. It turns out elegance is not the same as correctness. The next model will be more complex, and complexity is not a defect in science. It is often a sign that you are getting closer to the truth.

What happens next

The immediate future of cosmology looks messy. More DESI data is coming. Year three and year four observations will double the sample size and tighten constraints significantly. The Euclid space telescope, launched by the European Space Agency, is mapping the universe in three dimensions and will test these findings independently. The Nancy Grace Roman Space Telescope, set to launch soon, will add its own measurements to the mix. Each new dataset either tightens the constraint on evolving dark energy or reinforces the constant model. One of them will likely prove decisive within the next five years.

Beyond the data, there is a deeper shift underway. Cosmology is moving from a discipline that fits parameters to one that confronts model failure. The ΛCDM model was never intended to be eternal. It was always a placeholder — the best model we had, not the final model. That distinction matters. It means the current crisis is not a failure of science. It is science working. The discomfort physicists feel is the sensation of a boundary being reached, which is always followed by either refinement or revolution.

What is clear now is that the era of comfortable certainty is over. The standard model of cosmology survived because it worked well enough. It predicted the cosmic microwave background spectrum, the distribution of galaxies, the abundance of light elements. It made bold predictions that were confirmed. It was never meant to be final. It was always a work in progress.

The question now is whether the cracks are cosmetic or structural. If dark energy is truly evolving, the math changes. The geometry of spacetime changes. The story of the universe changes from a straightforward acceleration to something more complex — perhaps a cycle, perhaps a transition, perhaps something no one has fully articulated yet. We are not close to answering that question. But we are closer than we were a year ago, and that is more than can be said for most problems in physics.

The universe, it seems, is not done surprising us.