What is dark energy?

Gravity draws matter together. Why, then, can the separation of distant galaxies accelerate on the largest scales? The answer is unlike a wind pushing galaxies through space.

Dark energy is the name of a component used in cosmological models based on general relativity to explain the Universe’s accelerated expansion. Its physical nature remains unknown. The simplest option is a cosmological constant; evolving dark energy and modifications of gravity remain under investigation. PDG: the evidence and the assumptions behind its interpretation, §28.1–28.2.

The distance gained in equal times increases

In the large-scale description, the scale factor describes how distances between points following the overall expansion change with cosmic time. Acceleration means its increments over equal time intervals grow. This describes the average cosmic picture; an individual galaxy’s motion inside a cluster can differ.

The Hubble parameter describes a fractional rate: the increase in distance per unit time, divided by the distance itself. This is a different quantity. With a constant fractional rate, a larger distance gains more in absolute terms. Expansion can therefore accelerate even though the Hubble parameter is not increasing.

Constant fractional rate, larger increments

Start
Relative distance D/D₀: 1
After one interval
Relative distance D/D₀: 1.5
After two intervals
Relative distance D/D₀: 2.25

Equal intervals: +0.5, then +0.75

Each step adds 50% of the distance already present.

Hypothetical exponential example: D(t)=D₀·1.5ᵗ⁄ᵀ for points following the overall expansion; T is an equal positive interval. D₀ is the initial distance. Bars share a linear scale: full length =2.25D₀. These are not observations, a reconstruction of our Universe or the stretching of bound objects. Constant Hubble parameter H=ln(1.5)/T coexists with positive acceleration.

The example does not prove that our Universe has always expanded this way. It separates two questions: how quickly the scale grows and what percentage it gains per unit time. PDG: H and the cosmological equations, §22.1.3–22.1.5.

Where the evidence for acceleration came from

A telescope did not record billions of years of expansion in a single movie. Astronomers compare light’s redshift with distance estimated another way. Type Ia supernovae, corrected for light-curve shape and colour, serve as standardized luminosity indicators. Their apparent brightness helps determine luminosity distance. How the cosmic distance ladder works.

In papers by two teams in 1998–1999, distant supernovae were on average fainter than expected in the corresponding models without a cosmological constant. Within that comparison, the observations implied greater distances and supported an accelerating expansion history. Dust, sample selection and possible supernova evolution were examined; a single faint object cannot automatically be declared evidence for dark energy. Riess and colleagues, 1998, Perlmutter and colleagues, 1999.

Another tool is baryon acoustic oscillations, or BAO: a characteristic scale in the statistical distribution of matter, inherited from early sound waves. Its apparent size in different directions and at different redshifts helps reconstruct the expansion history. Combining this with the cosmic microwave background and other observations provides tests based on different physics. PDG: observational methods, §28.3.

Pressure also matters in general relativity

Ordinary matter with little pressure slows cosmic expansion. But in the equations of general relativity, changes in the scale depend on pressure as well as energy density. A component with sufficiently negative pressure can promote acceleration if its contribution dominates.

For a cosmological constant, the energy density stays constant and the pressure is negative, with a magnitude equal to that energy density. This is a property of the model, rather than an ordinary gas with a negative number of molecules. No external pump or edge of the Universe pulling space is required. PDG: negative pressure and the cosmological constant, §28.1.

Branch · the equationOptional · you can skip this

For a homogeneous and isotropic model in general relativity, write:

\frac{\ddot a}{a}=-\frac{4\pi G}{3c^2}(\varepsilon+3p).

Here $a$ is the dimensionless scale factor; dots denote derivatives with respect to cosmic time in seconds; $G$ is the gravitational constant and $c$ is the speed of light. The total energy density $\varepsilon$ is measured in J/m³ and total pressure $p$ in Pa, also J/m³. A cosmological constant, if present, is already included in these totals; do not add it a second time as a separate term.

Positive acceleration requires $\varepsilon+3p<0$ . For an individual component with positive energy density, define the dimensionless $w=p/\varepsilon$ . A value $w<-1/3$ contributes toward acceleration; it does not guarantee acceleration of the entire mixture until that contribution outweighs the others. For a cosmological constant, $w=-1$ . The form with explicit $c$ is obtained from PDG, equation 22.9, which uses units with $c=1$ .

What the name has not explained

Dark energy and dark matter play different roles in models. Dark matter helps explain the gravitational clustering of matter; dark energy describes the late accelerated change in the cosmic scale. “Dark” does not make them the same substance and does not specify a colour.

A cosmological constant can be described as a constant vacuum energy. However, relating the observed value to quantum fields remains a major theoretical problem. A successful description of expansion is not an established microscopic mechanism. And if gravity itself needs modification on cosmic scales, acceleration need not require a new energy component in precisely this form. PDG: dark energy or modified gravity, §28.2.1.

Branch · research status on 30 October 2026Optional · you can skip this

Has DESI proved that dark energy changes?

The 2025 DESI DR2 results combined BAO, the cosmic microwave background and different supernova samples. They preferred a model with an evolving equation of state at 2.8–4.2 σ. This depended on the supernova sample and the models being compared; it was not a direct measurement of an unknown substance “fading away.” DESI DR2 II, §VII and Table V.

A newer DESI preprint, presented on 30 July 2026, adds information from the shape of correlations in intergalactic hydrogen absorption. For a two-parameter evolving equation-of-state model, Table 4 gives 2.7 σ for DESI plus the cosmic microwave background and 3.2 σ after adding DES-Dovekie supernovae. These are particular data combinations, rather than a universal assessment of every observation. DESI DR2 IV, version 3: §VI.1, §VI.4.1 and Table 4.

The hint deserves investigation, but dark energy’s physical nature is not established, and its evolution cannot yet be presented as a discovered fact. Calibration, systematic uncertainties and the chosen description of evolution remain important. Completing a survey’s observations does not itself complete every analysis.

We measure expansion and continue testing its explanation

The evidence for acceleration comes from comparing light, distances and cosmic structure across epochs. Dark energy allows this to be described in general-relativistic models; the cosmological constant is the simplest candidate, but the name does not explain the phenomenon’s origin.

“The Universe’s expansion is accelerating” and “we know what dark energy is” therefore have different status. Predicting the distant future requires an additional assumption about how this component behaves beyond the history we have observed.