Current precision-cosmological data — Baryon Acoustic Oscillation (BAO) measurements from DESI and SDSS, Type Ia supernova (SN) compilations (Pantheon+, Union3, DESY5), cosmic microwave background (CMB) anisotropies, and cosmic-chronometer H(z) measurements — show a preference of roughly 2.5–4.2σ for an evolving dark energy equation of state in the Chevallier–Polarski–Linder (CPL) parametrization, whose strength depends strongly on which datasets are combined. The problem is to determine the origin of this dataset dependence: genuine late-time cosmological dynamics versus residual survey-specific systematics (SN photometric calibration and host-galaxy standardization, BAO tracer-level systematics and low-redshift coverage), and to establish whether the preference is robust across dark-energy parametrizations, statistical paradigms, and future data releases. As of August 2026 the question is open: recalibrated SN samples (DES-Dovekie, Union3.1) have narrowed the compilation dependence but leave a persistent ~3.2–3.4σ fit-level preference without decisive Bayesian evidence, and no identified systematic accounts for the full pattern.

Background

In the standard cosmological model (ΛCDM), the observed late-time acceleration of the Universe is driven by a cosmological constant Λ, a component with constant energy density and pressure-to-density ratio (equation of state, EoS) w=−1w=-1. Dark energy (DE) generalizations allow w≠−1w\neq -1 or a time-varying EoS w(a)w(a), where aa is the scale factor normalized to 1 today. The workhorse two-parameter description is the Chevallier–Polarski–Linder (CPL) parametrization, w(a)=w0+wa(1−a)w(a)=w_0+w_a(1-a), a linear Taylor expansion about a=1a=1: w0w_0 is the present-day EoS and waw_a its evolution rate; w0>−1w_0>-1 corresponds to quintessence-like behavior today and wa<0w_a<0 to a phantom-like (w<−1w<-1) regime in the past. Three observational pillars constrain this sector. (1) Baryon Acoustic Oscillations (BAO): the frozen acoustic peak in the pre-recombination baryon–radiation plasma provides a standard ruler for distance–redshift measurements (the sound horizon, the comoving distance sound waves could travel before recombination); the Dark Energy Spectroscopic Instrument (DESI) has produced the most precise BAO dataset to date (Data Release 1 in 2024, DR2 from three years of data in 2025), complemented by the earlier SDSS survey. (2) Type Ia supernovae (SNe Ia): approximately standardizable candles whose distance moduli are calibrated through light-curve-shape and color corrections plus host-galaxy properties; the main compilations are Pantheon+, Union3, and DESY5 (the five-year Dark Energy Survey SN sample), which differ in sample composition, calibration, and standardization methodology. (3) The cosmic microwave background (CMB): Planck (and ground-based ACT/SPT) anisotropies constrain the early-Universe physics anchoring the BAO ruler. A fourth, weaker probe is cosmic chronometers (CC): H(z) measurements inferred from the relative ages of massive, passively evolving galaxies, largely model-independent but statistically limited. In 2024–2025, DESI BAO combined with CMB and SNe was reported to favor the CPL point (w0,wa)(w_0,w_a) over (−1,0)(-1,0) — the cosmological constant — at 2.5–3.9σ in the DR1 era and 3.1σ (DESI BAO+CMB) to 2.8–4.2σ (with SNe, depending on the SN compilation) in DR2. The source review (Giarè, Mahassen, Di Valentino & Pan 2025) systematically examined 35+ dataset combinations and found the preference contingent on dataset choice: the only combination that significantly weakens it is SDSS BAO together with Pantheon+ SNe, while DESY5/Union3 SNe and DESI BAO strengthen it. Throughout, statistical preference is quantified both by Gaussian-equivalent significance (from Δχ2\Delta\chi^2 or likelihood-ratio tests) and by Bayesian model comparison, where the Bayes factor (the ratio of marginal likelihoods of two models, often inconclusive even when Δχ2\Delta\chi^2 is sizable) sets a stricter evidential bar. Whether these coherent, dataset-dependent shifts in (w0,wa)(w_0,w_a) are the first evidence of dynamical dark energy or the imprint of residual systematics in specific surveys is currently one of the most actively debated questions in cosmology.

Problem Statement

Determine the origin of the dataset-dependent preference for evolving dark energy in current precision-cosmological data — specifically, whether the stronger preferences associated with DESI BAO and with the DESY5/Union3 supernova compilations (2.5–3.9σ in the source's DR1-era analysis; 3.1σ for DESI BAO+CMB and 2.8–4.2σ with SNe in DESI DR2) reflect genuine late-time cosmological dynamics or residual survey-specific systematics — and establish whether the preference is robust. The problem separates into four linked strands posed by the source: (i) attribution of the dataset dependence — which specific systematics (SN photometric cross-calibration, host-galaxy standardization, low-redshift sample composition; BAO tracer-level fluctuations and survey redshift coverage) do or do not reproduce the observed shifts in (w0,wa)(w_0,w_a) across data combinations, versus a coherent cosmological signal; (ii) robustness of the preference's significance across dark-energy parametrizations and model classes beyond the CPL linear Taylor expansion of w(a)w(a) about a=1a=1; (iii) the constraining role and systematics control of cosmic-chronometer H(z) measurements, including combinations such as DESI+PantheonPlus where they materially alter the inference; and (iv) persistence of the preference under independent cross-checks and the updated data releases currently available (DESI DR2, reprocessed SN samples, independent BAO such as DES Year 6). The question is settled by establishing one of three scientific outcomes: (A) identified and quantified systematics that reproduce the dataset-dependent shifts in (w0,wa)(w_0,w_a) across the DESI/SDSS BAO and Pantheon+/Union3/DESY5 combinations; (B) a demonstrated cosmological signal that survives such systematics corrections and parametrization changes at decisive statistical evidence, defined as a Bayes factor of at least 10 (strong on the Jeffreys scale); or (C) a documented demonstration that, with the currently available releases, neither (A) nor (B) is attainable — specifying which datasets and systematics control the limit. This settles whether the hints of evolving dark energy are new physics, data/analysis artifacts, or not yet decidable with the current data.

References

  1. William Giarè, Tariq Mahassen, Eleonora Di Valentino, Supriya Pan, 'An overview of what current data can (and cannot yet) say about evolving dark energy', Physics of the Dark Universe 48 (2025) 101906, DOI 10.1016/j.dark.2025.101906, arXiv:2502.10264, https://arxiv.org/abs/2502.10264
  2. DESI Collaboration (M. Abdul-Karim, J. Aguilar, S. Ahlen, S. Alam, et al.), 'DESI DR2 Results II: Measurements of Baryon Acoustic Oscillations and Cosmological Constraints' (2025), arXiv:2503.14738, https://arxiv.org/abs/2503.14738
  3. Lu Huang, Rong-Gen Cai, Shao-Jiang Wang, 'The DESI DR1/DR2 evidence for dynamical dark energy is biased by low-redshift supernovae' (2025), arXiv:2502.04212, https://arxiv.org/abs/2502.04212
  4. Eoin Ó Colgáin, Saeed Pourojaghi, M. M. Sheikh-Jabbari, Lu Yin, 'How much has DESI dark energy evolved since DR1?' (2025), arXiv:2504.04417, https://arxiv.org/abs/2504.04417
  5. B. Popovic, P. Shah, W. D. Kenworthy, R. Kessler, T. M. Davis, A. Goobar, D. Scolnic, M. Vincenzi, et al., 'The Dark Energy Survey Supernova Program: A Reanalysis Of Cosmology Results And Evidence For Evolving Dark Energy With An Updated Type Ia Supernova Calibration' (2025), arXiv:2511.07517, https://arxiv.org/abs/2511.07517
  6. Taylor J. Hoyt, David Rubin, Greg Aldering, Saul Perlmutter, Andrei Cuceu, Ravi Gupta, 'Union3.1: Self-consistent Measurements of Host Galaxy Properties for 2000 Type Ia Supernovae' (2026), arXiv:2601.19424, https://arxiv.org/abs/2601.19424
  7. David Rubin, Taylor Hoyt, Greg Aldering, Saul Perlmutter, 'Banana Split: Improved Cosmological Constraints with Two Light-Curve-Shape and Color Populations Using Union3.1+UNITY1.8' (2026), arXiv:2601.19854, https://arxiv.org/abs/2601.19854
  8. Laura Herold, Tanvi Karwal, 'Bayesian and frequentist perspectives agree on dynamical dark energy' (2025), arXiv:2506.12004, https://arxiv.org/abs/2506.12004
  9. Mustapha Ishak, Leonel Medina-Varela, 'Persistent and serious challenge to the ΛCDM throne: Evidence for dynamical dark energy rising from combinations of different types of datasets' (2025), arXiv:2507.22856, https://arxiv.org/abs/2507.22856
  10. Mikel Artola, Ismael Ayuso, Ruth Lazkoz, Vincenzo Salzano, 'Is Chevallier-Polarski-Linder dark energy a mirage?' (2025), arXiv:2510.04191, https://arxiv.org/abs/2510.04191
  11. Tian-Nuo Li, Guo-Hong Du, Sheng-Han Zhou, Yun-He Li, Jing-Fei Zhang, Xin Zhang, 'Robust evidence for dynamical dark energy in light of DESI DR2 and joint ACT, SPT, and Planck data' (2025), arXiv:2511.22512, https://arxiv.org/abs/2511.22512
  12. Seokcheon Lee, 'Controlled Tension Forecasting: Quantifying Cross-Probe Biases in ω0ωa\omega_0\omega_aCDM' (2025), arXiv:2512.04130, https://arxiv.org/abs/2512.04130
  13. Hanyu Cheng, Supriya Pan, Eleonora Di Valentino, 'Beyond Two Parameters: Revisiting Dark Energy with the Latest Cosmic Probes' (2025), arXiv:2512.09866, https://arxiv.org/abs/2512.09866
  14. Hyeok Woo, William Luke Matthewson, Arman Shafieloo, 'Model-independent consistency tests of DESI DR2 BAO and SN Ia' (2026), arXiv:2604.19393, https://arxiv.org/abs/2604.19393
  15. Hayley J. Macpherson, Georgios Valogiannis, 'On the potential for inhomogeneities to mimic an evolving dark energy' (2026), arXiv:2606.05479, https://arxiv.org/abs/2606.05479
  16. Tengpeng Xu, Suresh Kumar, Yun Chen, Abraão J. S. Capistrano, Özgür Akarsu, 'Probing dynamical dark energy with late-time data: Evidence, tensions, and the limits of the w0waw_0w_aCDM framework' (2026), arXiv:2602.11936, https://arxiv.org/abs/2602.11936
  17. Anna Chiara Ferri, Ruchika, Alessandro Melchiorri, 'Present Day Cosmic Acceleration from SDSS and DESI BAO: A Call for Finer Tomography of the DESI Bright Galaxy Survey' (2026), arXiv:2607.07348, https://arxiv.org/abs/2607.07348
  18. Anil Kandel, Farruh Atamurotov, Ertan Güdekli, Phongpichit Channuie, Gulzoda Rakhimova, 'Revisiting the evidence for dynamical dark energy: from DES-SN5Y to DES-Dovekie', The European Physical Journal C 86 (2026) 762, DOI 10.1140/epjc/s10052-026-16010-6, https://doi.org/10.1140/epjc/s10052-026-16010-6
  19. Jose Agustin Lozano Torres, 'Dynamical Dark Energy from DESI DR2 and DES Year 6 BAO, Supernovae and Cosmic Chronometers: A 3σ Preference Without Decisive Bayesian Evidence', Monthly Notices of the Royal Astronomical Society (2026), DOI 10.1093/mnras/stag1607, https://doi.org/10.1093/mnras/stag1607