Dark energy may be a flaw in the model

July 22, 2026

The universe expands, and that expansion appears to accelerate. We have good evidence for both claims. We do not know why the second one is true.

The usual answer is dark energy. It is the name given to whatever drives the acceleration. In the standard account, it fills empty space and makes up about 70 percent of the energy in the universe. Yet no experiment has identified what it is.

A new mathematical result offers another answer. Perhaps there is no extra substance or force. Perhaps the acceleration follows from an instability in the simple model used to describe the universe.

That is a large claim, but also a narrow one. The paper proves a result about a class of equations. It does not prove that dark energy is absent from the real universe. Turning the result into a full account of what astronomers see will take much more work.

Why dark energy entered the picture

In the late 1990s, two teams studied distant Type Ia supernovae. These exploding stars have similar known brightness. Astronomers can compare their true brightness with how faint they look from Earth. That gives an estimate of distance.

The distant supernovae looked fainter than expected. The simplest reading was that cosmic expansion had sped up while their light travelled toward us.

The finding did not come from watching a galaxy move faster in real time. It came from fitting observations to models of cosmic history. Many later measurements supported an accelerating universe. These include the cosmic microwave background and patterns called baryon acoustic oscillations.

The standard Lambda-CDM model explains the acceleration with Lambda, Einstein's cosmological constant. Lambda gives empty space a fixed energy density. As the universe grows and matter spreads out, this constant part comes to dominate.

Lambda-CDM fits a wide range of data with few parameters. That success matters. But the model does not tell us what dark energy is. It also creates a severe mismatch between the observed value of Lambda and estimates from quantum field theory.

Recent results add another doubt. When scientists combine three years of DESI data with supernovae and other observations, models with changing dark energy can fit better than a fixed cosmological constant. The evidence ranges from 2.8 to 4.2 sigma, depending on the data used. That falls short of the five sigma standard for a discovery. DESI data alone still agree with Lambda-CDM.

So we should separate three statements:

The universe appears to expand at an increasing rate.
Lambda-CDM gives a strong fit to many observations.
A new form of energy causes the acceleration.

The first has firm evidence. The second is a fact about a model. The third remains an explanation, not a direct observation.

The hidden assumption

Modern cosmology starts with solutions to general relativity found by Alexander Friedmann. The related Friedmann-Lemaître-Robertson-Walker model treats the universe as uniform and directionless at large scales.

This is a useful simplification. The real universe contains galaxies, clusters, filaments and vast empty regions. Averaged over a large enough volume, however, it looks much smoother. The cosmological principle assumes that no place or direction is special at that scale.

The simplification makes Einstein's equations manageable. It also fixes the range of answers those equations can give. If the chosen solution is unstable, small departures from perfect uniformity may grow. The real universe could then follow a nearby solution with different behavior.

Christopher Alexander, Blake Temple and Zeke Vogler studied this possibility. They combined Einstein's equations for gravity with the equations for a pressureless fluid. They then rewrote them for spherical, self-similar expansion.

In this form, the standard flat Friedmann universe becomes a fixed point in a changing system. The authors studied what happens after small radial changes near the Big Bang.

A possible solution without dark energy

The paper proves that the flat Friedmann solution is unstable under the radial changes allowed by its setup. Curved Friedmann solutions are unstable too. Small underdense changes can push the system into a wider family of solutions.

Those solutions share three useful traits. They begin close to a Friedmann universe at leading order. At intermediate times, their expansion speeds up relative to that Friedmann background. At late times and at any fixed radius, they approach the background again.

This temporary extra acceleration comes from the original equations. The authors do not add a cosmological constant or a new field. Initial conditions select a path that naturally moves away from the ideal Friedmann solution.

That changes the question. Instead of asking what unseen energy pushes space apart, we can ask whether the apparent push comes from comparing an unstable, uneven universe with an ideal, uniform one.

The paper also argues that its free initial values can reproduce the coefficients in a series for luminosity distance against redshift. In principle, this gives the model enough room to mimic the distance pattern linked to dark energy, including a changing rate of acceleration.

The idea has an appealing economy. General relativity already allows the effect. The unknown sits in the initial shape of cosmic expansion instead of a new substance that fills space.

What the proof does not show

Mathematical instability does not by itself make a physical model false. A system can remain close to an unstable solution for a long time. The size, kind and growth of real perturbations all matter.

The proof also uses strong limits. It assumes spherical symmetry, radial changes, self-similar variables and pressureless matter. The authors focus on the matter-led era. They have not yet given the same full stability study for the earlier radiation-led era.

Most important, the proposed solutions have a center. To see nearly the same expansion in every direction, the Milky Way would need to lie near the right part of that structure. This weakens the usual claim that we occupy no special cosmic location. Past void models faced tight limits for the same reason.

A working replacement for Lambda-CDM must fit much more than supernova distances. It must also explain the microwave background, galaxy growth, gravitational lensing, element formation and acoustic patterns. It should do so with one set of initial conditions, not a new adjustment for each test.

The authors present a mechanism and a mathematical family of possible universes. They do not yet present a single best-fit universe tested against all major data sets.

The test that matters

The next step is numerical and observational. Researchers need to evolve these unstable solutions through radiation and matter eras. They must then calculate clear predictions for several independent measurements.

A useful model should predict where it differs from Lambda-CDM. It might imply a specific relation between distance, redshift and the growth of galaxies. It might leave a distinct mark on the microwave background. It might also fail quickly once placed against those data.

Either outcome would help. If the model fails, it sharpens the case for dark energy or another change to gravity. If it succeeds, dark energy may turn out to be a correction for a model that was too smooth from the start.

The new paper does not take dark energy out of cosmology. It shows a path by which Einstein's old equations may produce the same apparent effect. That possibility deserves tests, not a victory lap.

References

#physics