DESI DR2 Analysis Finds 3 Sigma Evidence for Dynamical Dark Energy After DES-SN5YR Recalibration

The study reexamined five dark energy models including CPL, BA, EXP, LOG and JBP using DESI DR2 baryon acoustic oscillation measurements, Planck and ACT CMB power spectra, and three supernova samples. With the DES Collaboration's Dovekie recalibration of DES-SN5YR, the highest previous 4.2 sigma signal dropped to 3.2 sigma, while Pantheon+ and Union3 produced 2.1-3.0 sigma results. All models consistently preferred a dark energy equation of state that was phantom-like at higher redshifts and transitioned to quintessence-like behavior by the present epoch.

This pattern challenges the cosmological constant assumption in Lambda CDM and aligns with earlier hints from DESI DR1 but survives the key systematic correction in supernova photometry. The 3 sigma level remains below the 5 sigma discovery threshold yet indicates that the tension is not driven solely by the original DES-SN5YR calibration error. Cross-checks with ACT DR6 CMB data further reduce the chance that the signal arises from Planck-specific foreground modeling.

Future DESI year-3 and year-5 releases will increase BAO precision by roughly 40 percent, while LSST and Roman supernova samples should halve current calibration uncertainties. If the dynamical signal persists above 4 sigma in these larger datasets by 2028, it would force revisions to standard inflation and late-time acceleration models.

The main limitation is residual supernova calibration uncertainty at the 0.02-0.03 mag level; independent photometric recalibrations from multiple surveys would strengthen the evidence.