A fundamental question in cosmology is whether dark energy evolves over time, a topic that has gained prominence since the discovery of cosmic acceleration. Recently, the DESI collaboration has reported increasing evidence for evolving dark energy using combinations of cosmic microwave background (CMB), type Ia supernova (SN), and their new measurements of baryon acoustic oscillations (BAO). However, our analysis reveals that these combinations are problematic due to clear tensions among the CMB, BAO and SN datasets. Consequently, DESI's claim of dynamical dark energy (DDE) is not robust. A more reliable approach involves constraining the evolution of dark energy using each dataset independently. Through a statistical comparison for each dataset, on average, we find that DDE is strongly preferred over the $Î$CDM model. This suggests that DDE likely exists, although its real parameter space remains elusive due to weak constraints on the dark energy equation of state and inconsistencies among the datasets. Interestingly, when considering DDE, none of the individual datasets -- including CMB, DESI DR2, Pantheon+, Union3, and DESY5 -- can independently detect cosmic acceleration at a significant level. Our findings not only clarify the current understanding of the nature of dark energy but also challenge the established discovery of cosmic acceleration and the long-held notion that dark energy exerts negative pressure. Both individual and combined datasets suggest that the ultimate fate of the universe is likely to be dominated by matter rather than dark energy.
- Paper ID: 2504.15222
- Title: Did DESI DR2 truly reveal dynamical dark energy?
- Authors: Deng Wang (Instituto de Física Corpuscular, CSIC-Universitat de València), David Mota (University of Oslo)
- Classification: astro-ph.CO (Cosmology and Non-Galactic Astrophysics), astro-ph.HE, gr-qc
- Publication Date: November 20, 2025 (arXiv v2)
- Paper Link: https://arxiv.org/abs/2504.15222
This paper questions DESI DR2's claims regarding dynamical dark energy (DDE). The authors point out that the DDE evidence obtained by the DESI collaboration through combining cosmic microwave background (CMB), Type Ia supernovae (SN), and baryon acoustic oscillation (BAO) data is problematic due to apparent tensions between these datasets. Through independent analysis of each dataset, the authors find that while DDE is statistically favored over the ΛCDM model, the true parameter space is difficult to determine. More strikingly, when considering DDE, no single dataset can independently detect cosmic acceleration at a significant level, challenging the established finding of cosmic acceleration and the long-held notion of dark energy's negative pressure. The research suggests that the universe's ultimate fate may be dominated by matter rather than dark energy.
Whether dark energy evolves with time is a fundamental question in cosmology. The DESI collaboration claims to have discovered evidence for dynamical dark energy at the 2.8-4.2σ level through DR2 data, which would have revolutionary implications for theoretical physics.
- Theoretical Significance: If dark energy truly evolves, it implies the vacuum is non-empty and contains matter, potentially overturning existing cosmological theories
- Cosmic Destiny: DDE would significantly affect the ultimate fate of the universe
- Standard Model Challenges: The ΛCDM model faces the cosmological constant problem, coincidence problem, and tensions in H₀ and S₈ measurements
- DESI's DDE evidence derives from data combination rather than independent verification by a single probe
- Insufficient examination of consistency between CMB, BAO, and SN datasets
- Data combination may mask internal tensions, leading to spurious signals
The authors argue that one should:
- Rigorously examine tensions between datasets
- Independently assess each dataset's support for DDE
- Employ more conservative methods to confirm DDE's existence
- Revealing Data Tensions: Systematically identifies Ωₘ and H₀ tensions between CMB, DESI DR2, and three SN samples (Pantheon+, Union3, DESY5), with tension levels ranging from 1.5σ to 5.3σ
- Questioning Combined Analysis: Demonstrates that DESI's claimed DDE evidence is problematic because it relies on data combinations with significant internal inconsistencies
- Independent Statistical Tests: Through Bayesian evidence and BIC criteria, finds that individual datasets statistically favor DDE, but cannot determine the true parameter space
- Challenging Cosmic Acceleration: Discovers that within the DDE framework, no single dataset can independently detect cosmic acceleration at a significant level, challenging the Nobel Prize-winning discovery of 1998
- Cosmic Fate Prediction: Proposes that the universe's ultimate fate may be matter-dominated (Ωₘ→1) rather than dark energy-dominated, with the universe potentially ceasing to expand
Constraining the dark energy equation of state (EoS) within the Chevallier-Polarski-Linder (CPL) parameterization framework:
- Input: CMB power spectrum, BAO measurements, SN distance moduli
- Output: Posterior distributions of cosmological parameters, particularly (ω₀, ωₐ, Ωₘ, H₀)
- Constraint: ω₀ + ωₐ < 0 (ensuring matter domination at high redshift)
CPL dark energy model dimensionless Hubble parameter:
E(a)=[Ωma−3+(1−Ωm)a−3(1+ω0+ωa)e3ωa(a−1)]1/2
Where:
- ω₀: current dark energy equation of state
- ωₐ: dark energy evolution amplitude
- When ω₀ = -1, ωₐ = 0, it reduces to ΛCDM
- CMB: Planck 2018 high-ℓ TT/TE/EE + low-ℓ Commander + lensing potential
- BAO: 13 measurements from DESI DR2 (zₑff = 0.295-2.33)
- SN: Three calibrated samples
- Pantheon+: 1701 data points (z ∈ 0.00122, 2.26137)
- Union3: 22 spline interpolation points (z ∈ 0.05, 2.26)
- DESY5: 1735 effective points (z ∈ 0.025, 1.130)
1. MCMC Bayesian Inference
- CAMB used to compute theoretical power spectra
- Cobaya implements MCMC sampling
- Gelman-Rubin convergence criterion: R-1 ≲ 0.01
2. Prior Settings
Employing broad priors to fully explore parameter space:
- ω₀ ∈ -15, 20
- ωₐ ∈ -30, 10
- Other standard cosmological parameters use conventional priors
3. Statistical Comparison
- CMB: Bayes factor lnB_ij = ln(ε_CPL/ε_ΛCDM)
- lnB > 5: strong evidence
- 2.5 < lnB < 5: moderate evidence
- 1 < lnB < 2.5: weak evidence
- BAO/SN: Bayesian Information Criterion ΔBIC = BIC_CPL - BIC_ΛCDM
- ΔBIC ~ 6: strong evidence
- ΔBIC ~ 2: positive evidence
- Systematic Tension Analysis: First comprehensive quantification of multi-dimensional tensions between CMB, BAO, and SN under both ΛCDM and CPL models
- Independent Constraint Strategy: Emphasizes the importance of independent single-dataset constraints rather than blind combination
- Broad Prior Methodology: Uses broader priors than Planck and DESI, fully presenting the data's constraining power
- Multiple Calibration Tests: Detailed discussion of SN calibration and marginalization effects in supplementary materials
CMB Data:
- Planck 2018 TT: 30 ≤ ℓ ≤ 2508
- EE/TE: 30 ≤ ℓ ≤ 1996
- Low-ℓ: 2 ≤ ℓ ≤ 29
- Lensing potential: 8 ≤ ℓ ≤ 400 (SMICA maps)
DESI DR2 BAO:
13 measurements across 7 redshift bins, including BGS, LRG1-3, ELG1-2, QSO, and Lyα samples
SN Sample Selection Rationale:
- Using calibrated SN data to improve precision and reduce systematic errors
- Three samples provide cross-validation and systematic error testing
- Parameter Constraints: Posterior distribution means and 1σ/2σ confidence intervals
- Tension Level: Standard deviation multiples of posterior means between two datasets
- Statistical Significance:
- Bayes factor (CMB)
- ΔBIC (BAO/SN)
- Physical Quantities: rd (sound horizon), H₀rd composite parameter
- Baseline Model: ΛCDM (ω₀ = -1, ωₐ = 0)
- Extended Model: CPL parameterized DDE
- Data Combinations: Single dataset vs. pairwise combinations vs. all three combined
- GetDist used for MCMC chain analysis
- Strict convergence criterion (R-1 ≲ 0.01)
- SH0ES calibration applied to SN data
- Supplementary materials explore effects of lensing amplitude A_L
H₀ Tension (Figure 1 top):
- Pantheon+ vs CMB: 5.31σ
- DESY5 vs CMB: 3.03σ
- Union3 vs CMB: 1.69σ
- Pantheon+ vs DESY5: 3.86σ (internal SN inconsistency)
Ωₘ Tension:
- DESI DR2 vs CMB: 1.57σ
- SN samples favor larger Ωₘ
Tension Alleviation and New Problems (Figure 1 bottom):
- Ωₘ tension significantly relieved (reduced to 0.14σ-0.94σ)
- Cost: Pantheon+ vs DESY5 H₀ tension increases to 4.29σ
- DESI DR2 and SN give Ωₘ ~ 0.5, challenging the dark energy-dominated picture
Parameter Constraints (Table I):
| Dataset | ω₀ | ωₐ | Ωₘ | H₀ |
|---|
| CMB | 2.4±1.9 | < -3.62 | 0.344⁺⁰·⁰⁵⁴₋₀.₂₀₀ | Unconstrained |
| DESI | -0.17±0.44 | -2.8±1.6 | 0.385±0.049 | Unconstrained |
| Pantheon+ | -0.89±0.17 | -2.1⁺³·²₋₁.₈ | 0.360⁺⁰·¹³⁰₋₀.₀₈₆ | 73.2±1.0 |
| Union3 | -0.45⁺⁰·²⁸₋₀.₄₀ | -5.4⁺⁴·⁷₋₃.₁ | 0.437⁺⁰·¹⁰⁰₋₀.₀₆₆ | 71.1±3.0 |
| DESY5 | -0.35⁺⁰·³⁰₋₀.₄₁ | -9.0⁺⁵·⁴₋₄.₅ | 0.471⁺⁰·⁰⁷⁵₋₀.₀₄₃ | 68.20±0.60 |
rd Tension:
- DESI+Pantheon+ vs CMB: 5.30σ
- DESI+DESY5 vs CMB: 4.26σ
- DESI+Union3 vs CMB: 2.86σ
(ω₀, ωₐ) Parameter Space Tension:
- CMB vs DESI DR2: ~2σ
- CMB vs SN: ~1σ
- Pantheon+ vs DESY5+DESI: >2σ (internal SN inconsistency)
Key Finding: CMB+DESI combination shows 1.5σ-3σ tension with individual SN samples
DDE vs ΛCDM Preference:
- CMB: lnB = 6.34 (strong evidence)
- DESI DR2: ΔBIC = 2.25 (positive evidence)
- Pantheon+: ΔBIC = 25.63 (strong evidence)
- Union3: ΔBIC = 3.92 (positive evidence)
- DESY5: ΔBIC = 11.49 (strong evidence)
Conclusion: Individual datasets statistically favor DDE, but provide inconsistent parameter constraints
- When A_L is left free, DDE significance decreases by ~1σ
- Indicates degeneracy between A_L and DDE in CMB lensing potential
- Larger H₀ leads to smaller rd
- rd inherits the anti-correlation between H₀ and Ωₘ
- H₀rd = 91.5⁺⁴·⁴₋₄.₉ (DESI DR2 free fit)
- Calibrated Pantheon+ provides more conservative constraints
- Union3 and DESY5 constraints remain essentially consistent before and after calibration
- Justifies the use of calibrated data
Cosmic Acceleration Test:
In the CPL framework, cosmic acceleration requires ω₀ < 1/3(Ωₘ-1)
- CMB: ω₀ = 2.4±1.9, giving 1.3σ hint of ω₀ > 0 (positive pressure!)
- DESI DR2: ω₀ = -0.17±0.44, allowing ω₀ > 0 within 1σ
- Union3: ω₀ = -0.45⁺⁰·²⁸₋₀.₄₀, allowing ω₀ > 0 within 1σ
- DESY5: ω₀ = -0.35⁺⁰·³⁰₋₀.₄₁, allowing ω₀ > 0 within 1σ
- Pantheon+ alone: Provides >1σ hint of acceleration
Striking Conclusion: Within the DDE framework, the universe may currently be decelerating or moving uniformly, not accelerating!
- Spurious Signals from Data Combination: Combined constraints yield Ωₘ = 0.3108±0.0058, close to CMB-only ΛCDM's Ωₘ = 0.3153±0.0073, but the authors argue this is a coincidence
- Matter-Dominated Late Universe: SN data allow Ωₘ > 0.5, implying the late-time universe may be dominated by matter rather than dark energy
- Ultimate Cosmic Fate: Both independent and combined datasets favor Ωₘ → 1, suggesting the universe will ultimately be completely filled with matter, with expansion potentially miraculously ceasing
- Questioning Dark Energy Pressure: CMB allows dark energy to have positive pressure, overturning conventional wisdom
- Challenges to ΛCDM:
- Cosmological constant problem Weinberg 1989, Carroll 2001
- Coincidence problem
- H₀ tension and S₈ discrepancy Di Valentino et al. 2021-2022
- BAO Probes:
- History: 2dF, 6dF, SDSS, BOSS, eBOSS
- Current: DESI DR1/DR2 2024-2025
- Advantages of BAO as a "standard ruler" (~150 Mpc)
- DDE Theory:
- CPL parameterization Chevallier & Polarski 2001, Linder 2003
- Effects on cosmic evolution, structure formation, and ultimate fate
- DESI's DDE Claims:
- DR1: Initial evidence 2024
- DR2: Enhanced to 2.8-4.2σ 2025
Differences from DESI Official Analysis:
- DESI emphasizes significance of combined constraints
- This paper emphasizes independent constraints and data tensions
- DESI uses standard priors; this paper uses broad priors
Connection to Tension Research:
- Extends H₀ tension research to the DDE framework
- Reveals complex behavior of Ωₘ tension under CPL
- Systematicity: First comprehensive analysis of multi-dimensional tensions between CMB-BAO-SN in the DDE framework
- Independence: Emphasizes the importance of independent dataset verification
- Conservatism: Uses broad priors without presetting DDE parameter space
- Profundity: Challenges the fundamental discovery of cosmic acceleration
- DESI's DDE Claims Are Not Robust: Due to significant tensions (1.5σ-5.3σ) between CMB, BAO, and SN datasets, the DDE evidence based on data combination is problematic
- DDE May Exist But Parameters Are Uncertain: Individual datasets statistically favor DDE (lnB = 6.34, ΔBIC = 2.25-25.63), but provide inconsistent parameter constraints
- Cosmic Acceleration Is Questioned: Within the DDE framework, no single dataset can independently detect cosmic acceleration at a significant level; CMB even hints at positive pressure at 1.3σ
- Matter-Dominated Cosmic Fate: Both independent and combined datasets favor Ωₘ → 1, suggesting the universe will ultimately be matter-dominated, with expansion potentially ceasing
- Higher Precision Observations Needed: Only if future CMB, BAO, or SN observations provide consistent constraints can one safely claim DDE exists
- Model Dependence: Only considers CPL parameterization; other DDE models may yield different results
- Systematic Errors: While using calibrated SN data, complete treatment of systematic errors could be improved
- Prior Sensitivity: Broad priors, while demonstrating data constraining power, may introduce edge effects
- Statistical Methods: Tension quantification primarily uses standard deviation multiples; more sophisticated methods (e.g., tension measures) might yield different conclusions
- Physical Interpretation: Lacks in-depth discussion of physical mechanisms for dataset inconsistencies
- High-Precision Observations:
- Next-generation CMB experiments (CMB-S4, LiteBIRD)
- Complete 5-year DESI data
- New SN surveys (Rubin Observatory)
- Theoretical Development:
- Explore alternative DDE parameterizations
- Study modified gravity theories
- Understand physical origins of data tensions
- Methodological Improvements:
- Develop more robust multi-dataset combination methods
- Improve systematic error handling
- Establish unified framework for tension quantification
- Cosmic Fate Research: Authors will explore the universe's "Big Stall" in detail in follow-up work arXiv:2504.15635
- Problem Importance: Challenges DESI's major claims and the fundamental discovery of cosmic acceleration, with far-reaching implications
- Rigorous Analysis:
- Systematically quantifies multi-dimensional tensions
- Combines independent and joint analysis
- Employs multiple statistical tests (Bayes factors, BIC)
- Conservative Methodology:
- Broad prior settings avoid bias
- Emphasizes independent verification
- Detailed supplementary materials discuss systematic effects
- Striking Results:
- Questions cosmic acceleration (1998 Nobel Prize discovery)
- Challenges dark energy's negative pressure
- Predicts matter-dominated cosmic fate
- Clear Writing: Rigorous logic, abundant figures (3 main figures + 3 supplementary figures + 1 table)
- Risk of Over-Interpretation:
- Jump from "cannot significantly detect acceleration" to "universe may be decelerating" is substantial
- Weak constraints under broad priors may amplify uncertainties
- Insufficient Systematic Error Discussion:
- While discussing SN calibration, less emphasis on BAO systematic errors
- Incomplete discussion of theoretical modeling uncertainties (e.g., nonlinear effects)
- Missing Tension Explanation:
- Points out tensions but lacks deep exploration of physical origins
- Is it new physics or systematic errors? Lacks judgment
- Limited Statistical Methods:
- Tension quantification primarily uses standard deviation multiples
- Could supplement with other methods (e.g., suspicion statistics, tension measures)
- Limited Constructive Suggestions:
- Strong criticism but relatively few constructive recommendations
- How to improve data combination methods? How to design future observations?
- Lacking Physical Mechanisms:
- If the universe truly undergoes "Big Stall," what is the physical mechanism?
- What is the microscopic origin of DDE?
Contributions to the Field:
- Highly Important: May trigger re-examination of DDE evidence across cosmology
- Methodological Contribution: Emphasizes importance of independent verification and tension testing
- Theoretical Challenge: Forces reconsideration of cosmic acceleration and dark energy's nature
Practical Value:
- Provides important warning for DESI and other surveys
- Offers methodological guidance for future multi-dataset analysis
- Provides new research directions for theorists
Reproducibility:
- Uses public data and tools (Planck, DESI DR2, Pantheon+/Union3/DESY5)
- Detailed method description (CAMB, Cobaya, GetDist)
- Supplementary materials provide additional technical details
- High reproducibility
Potential Controversy:
- DESI collaboration may disagree with "not robust" conclusions
- Questioning cosmic acceleration may spark heated debate
- Results require independent team verification
- Cosmological Parameter Constraints: Any research involving CMB-BAO-SN combined analysis should consider data tensions
- DDE Model Verification: Provides important observational constraints and warnings for DDE theoretical research
- Future Survey Design: Guides how to design observations to resolve data inconsistencies
- Systematic Error Research: Emphasizes critical role of systematic errors in high-precision cosmology
- Scientific Methodology: Serves as exemplary case for emphasizing independent verification and conservative analysis
- Planck 2018: Aghanim et al., A&A 641, A6 (2020) - CMB baseline constraints
- DESI DR2: Abdul Karim et al., arXiv:2503.14738/14739/14743 (2025) - Subject of this paper's critique
- Pantheon+: Brout et al., ApJ 938, 110 (2022) - SN data
- CPL Parameterization: Chevallier & Polarski (2001), Linder (2003) - DDE theoretical framework
- Cosmological Tension Review: Di Valentino et al., Astropart. Phys. 131 (2021), JHEAp 34 (2022)
- Wang, "Questioning Cosmic Acceleration with DESI: The Big Stall of the Universe," arXiv:2504.15635 - Detailed exploration of cosmic fate
Overall Assessment: This is a highly important and controversial paper that rigorously questions DESI's major claims regarding dynamical dark energy and further challenges the fundamental discovery of cosmic acceleration. The methodology is conservative and analysis systematic, though some conclusions (such as the universe's "Big Stall") require additional supporting evidence. Regardless of final conclusions, this paper will promote the cosmology community's attention to data consistency and independent verification, possessing significant methodological value. Readers are advised to read critically and pay attention to subsequent independent verification and DESI collaboration's response.