The Hubble tension — a ~5σ discrepancy between local measurements of the cosmic expansion rate (H₀ ≈ 73 km/s/Mpc from Cepheids and supernovae) and early-universe measurements from the CMB (H₀ ≈ 67 km/s/Mpc from Planck) — has resisted resolution for fifteen years. We propose that the tension arises not from systematic error or unknown particle physics, but from a clock mismatch: CMB measurements use the identity clock, which encodes the thermodynamic history of the early universe; distance-ladder measurements use the geometry clock, which encodes spatial relationships in the late universe. These two clocks, identified in the Unified Fluid Field Theory (UFFT) as distinct temporal channels of the 12-dimensional substrate from which spacetime emerges, run at slightly different rates across cosmic time because the vacuum structure of the substrate evolves as the universe expands and cools. We derive a substrate correction to the Friedmann expansion history of the form H(z) → H(z)√[1 + ε_S f_S(z)], where f_S(z) is peaked in the pre-recombination era (z ~ 2000–5000). A parameter scan gives H₀ = 68.5–70.4 km/s/Mpc for ε_S = 0.12–0.35, reducing the Hubble tension from ~5σ to 2.3–4.0σ while preserving CMB acoustic peak positions, BBN abundances, BAO measurements, and gravitational wave propagation speed. For ε_S = 0.35, a predicted power deficit of ~8% at CMB multipole ℓ = 2000 provides a falsifiable signature detectable by CMB-S4. The same two-clock framework accounts for the R_mg particle physics measurements in Barber (2026a, DOI: 10.5281/zenodo.18821542) and the black hole information paradox resolution in Barber (2026b, DOI: 10.5281/zenodo.18821669). The universe has been trying to tell us it has two clocks for a century. We have been assuming it has one.