A collaboration between scientists in Malta and Istanbul has uncovered an intriguing signal suggesting that one of cosmology’s most mysterious components — dark energy — may not have remained constant throughout the history of the universe.
Instead, a data-driven reconstruction indicates that dark energy’s effective density could have been negative in the distant past before transitioning to positive values as the universe evolved.
The result would represent a major departure from the simplest version of the standard cosmological model if it survives future tests.
But researchers are emphasizing an essential distinction: this is not yet a discovery.
The evidence remains suggestive rather than conclusive, and further observations — particularly information connected to the early universe — will be needed before physicists can determine whether they are seeing new fundamental physics or a feature that disappears as measurements improve.
What Is Dark Energy?
Scientists discovered in the late 1990s that the expansion of the universe is accelerating rather than slowing down.
The unknown component used to explain that acceleration became known as dark energy.
One way to visualize cosmic expansion is to imagine dots drawn on the surface of an inflating balloon. As the balloon expands, the dots become farther apart even though they are not independently traveling across the balloon’s surface.
In a roughly analogous way, the distances between widely separated galaxies increase as space itself expands.
The standard Lambda Cold Dark Matter model, or ΛCDM, represents dark energy using the cosmological constant, commonly written as Λ. In its simplest form, this component has a constant energy density throughout cosmic history.
That remarkably simple model has successfully described a wide range of astronomical observations.
Yet cosmologists continue investigating whether dark energy might be more complicated, particularly as increasingly precise measurements expose tensions that are difficult to reconcile within one straightforward picture.
Malta and Istanbul Approach the Problem Differently
The new research brings together scientists led by Jackson Levi Said of the University of Malta and Özgür Akarsu of Istanbul Technical University (ITU).
Their approaches initially came from different directions.
Akarsu and collaborators had already explored theoretical cosmologies in which the effective cosmological constant could switch from negative to positive during the universe’s evolution.
A 2023 study involving Akarsu examined a model known as ΛsCDM, in which such a transition occurs at a redshift of roughly z ≈ 2. The model was investigated partly because of its potential implications for several persistent cosmological tensions.
The newer Malta-Istanbul project approached the problem from the observational side.
Rather than beginning by imposing a predetermined dark-energy evolution on the universe, the researchers reconstructed the late-time expansion history directly from observational data and then examined what kind of dark-energy behavior could reproduce it.
The Data Pointed Toward a Sign Change
The team used combinations of major cosmological data sets, including DESI and SDSS baryon acoustic oscillation measurements, Pantheon+ Type Ia supernovae and cosmic chronometers.
They also examined combinations involving external measurements of the present-day expansion rate.
The reconstruction uses a node-based Gaussian-process-kernel interpolant to estimate the universe’s expansion history without forcing it into one of the conventional low-dimensional dark-energy parameterizations.
When researchers translated that reconstructed expansion history into an effective dark-energy description under general relativity, something unusual appeared.
Across all of the data combinations examined in the February 2026 paper, the inferred effective dark-energy density crossed zero.
It was negative at higher redshifts and became positive toward the present-day universe.
That behavior is particularly interesting because it resembles a possibility that Akarsu and other theorists had explored before this reconstruction was performed.
A Possible Second Surprise Appeared
The sign change was not the only unusual feature.
Some combinations of observations produced hints of an additional, temporary period of accelerated cosmic expansion around redshifts of approximately z ≈ 1.7–2.3.
That corresponds to an intermediate stage of cosmic history rather than the acceleration dominating the universe today.
The researchers did not begin their analysis by requiring such a phase to exist.
Instead, the feature emerged from some of the reconstructed expansion histories.
This is precisely the kind of result that makes relatively assumption-light reconstruction techniques valuable: they can reveal structures that might be suppressed when researchers require the data to follow a predetermined mathematical form.
But that flexibility also requires caution.
A reconstruction capable of accommodating additional features can potentially fit fluctuations that do not ultimately represent fundamental physics.
Follow-Up Study Finds a 2–3 Sigma Tension
The team subsequently investigated whether conventional dark-energy parameterizations could reproduce the same behavior.
Their April 2026 analysis compared the reconstruction with several smooth, lower-dimensional descriptions of dark energy while using equivalent late-time observational combinations.
Both approaches produced broadly compatible expansion histories over the range directly constrained by the observations.
However, an interesting difference appeared around redshift z ≈ 1.7.
The reconstruction preferred substantially stronger cosmic deceleration in that region than the smoother dark-energy parameterizations, creating a persistent discrepancy of roughly 2–3 sigma depending on the data combination and model used.
In particle physics and cosmology, that level is intriguing but far below the conventional threshold required to claim a discovery.
There is another reason for caution.
Although the flexible reconstruction achieved improved best-fit likelihoods, the researchers reported that Bayesian evidence continued to favor the simpler parametric descriptions in their comparison.
That means the additional flexibility may describe the existing measurements better at particular points without yet providing sufficiently strong statistical evidence to establish a more complicated cosmological picture.
Could Dark Energy Really Change Sign?
If the result eventually survives independent analyses and increasingly precise observations, the implications could be profound.
A dark-energy density that moves from negative to positive would be fundamentally different from the fixed positive cosmological constant of standard ΛCDM.
It could imply that what scientists describe as the vacuum energy of the universe has a more complicated history than previously assumed.
The February analysis also examined what kinds of scalar fields could reproduce the reconstructed behavior.
The researchers found that a simple canonical scalar field cannot smoothly generate the inferred evolution, while alternatives involving a phantom field or a two-field “quintom” system could accommodate it.
That does not demonstrate that such fields actually exist.
Rather, it shows the kinds of theoretical ingredients that would be necessary if the reconstructed behavior reflects physical reality.
The Next Observations Will Be Crucial
The strongest tests are still ahead.
Late-time measurements alone cannot settle the issue. The proposed behavior must remain consistent with observations of the early universe, including the cosmic microwave background.
Future and expanded observations from projects including DESI and Euclid, together with cosmic microwave background measurements, can provide much tighter constraints on the universe’s expansion history.
The particularly interesting region around redshift 1.5–2 now represents an important target.
If increasingly precise observations continue producing the same feature, the argument for a changing dark-energy component would become considerably stronger.
If the signal fades, the cosmological constant could remain the simpler explanation.
A New Chapter in the Dark Energy Mystery
The Malta-Istanbul collaboration illustrates how modern cosmology is increasingly combining large astronomical surveys, flexible computational reconstruction techniques and theoretical physics.
The research team includes Özgür Akarsu, Maria Caruana, Konstantinos F. Dialektopoulos, Luis A. Escamilla, Emre O. Kahya and Jackson Levi Said. Their February 2026 study presents the result explicitly as evidence for possible sign-changing dark-energy density rather than as a confirmed detection.
The wider Bridging Cosmology project is led by Said at the University of Malta’s Institute for Space Sciences and Astronomy and Akarsu at Istanbul Technical University, with financing through the Xjenza Malta-TÜBİTAK Joint Call for R&I Proposals 2024, according to reporting on the collaboration.
For now, the safest conclusion is also the most scientifically interesting one: the standard cosmological constant has not been overturned, but current observations appear to leave room for a considerably stranger history of dark energy.
The universe may ultimately prove that its accelerated expansion is driven by something that has remained constant for billions of years.
Or the component responsible for that acceleration may itself have evolved — potentially even crossing from negative to positive energy density during cosmic history.
The data are not yet strong enough to choose between those possibilities. But the Malta-Istanbul team’s reconstruction has identified where future observations can look for the answer.