A multi-probe observational analysis has constrained a clustering dark energy model based on a tachyonic Dirac-Born-Infeld (DBI) action. Its noncanonical kinetic term yields a subluminal sound speed, allowing dark energy perturbations to grow on subhorizon scales, something ΛCDM does not permit. The standard cosmological parameters agree with ΛCDM within 2σ, with a slight preference for a lower H₀ and a higher Ωm0, the largest shift being ≈1.8σ in H₀ for the Pantheon+ compilation.

The team worked in a multifluid framework tracking both the background expansion and linear perturbations, and fitted the model using nested sampling. They combined Planck 2018 compressed CMB distance priors, DESI DR2 BAO data, cosmic chronometers, fσ₈ growth data and three Type Ia supernova compilations: Pantheon+, DES-SN5YR and Union3. Bayesian evidence favours the DBI model in all three compilations, moderately for Pantheon+ (ΔlnZ ≈ 3.4) and inconclusively for DES-SN5YR and Union3 (ΔlnZ ≈ 0.6-0.7).

The field thaws naturally, and dark energy perturbations remain several orders of magnitude below those of matter. Even so, the noncanonical sound speed alters the Hubble friction term and raises S₈ by only ≈0.5σ relative to ΛCDM. Four independent diagnostics —the CPL parametrization, Om(z), the statefinder hierarchy and the growth index— all point to genuinely evolving dark energy rather than a cosmological constant, while the early thermal history is left intact, with negligible early dark energy.

One structural limitation remains: since wφ = φ̇² − 1 ≥ −1, the phantom regime is out of reach, in line with the no-go theorem for a single minimally coupled scalar field. The DBI scenario therefore cannot follow the DESI DR2 preference for an effective equation of state crossing w = −1, a tension that future supernova and BAO data will have to clarify.