The Graviton Condensation Cosmology (GCC) framework posits that macroscopic gravity emerges from the condensation of spin-2 collective quadrupole excitations (GQR phonons) in nuclear matter. We extend this framework to address the dark sector: we propose that both dark matter and dark energy originate as residues of the QCD phase transition at T_c ≈ 156.5 MeV. Dark matter is identified with multi-glueball bound states — the unique GCC-compatible candidate after a systematic exclusion of all Standard Model and beyond-Standard-Model particles, which lack either QCD substructure or a spin-2 quadrupole channel. The relic abundance Ω_gb h² naturally matches the observed Ω_DM h² ≈ 0.12 for formation fractions f_gb ∼ (1—10) × 10u207bu2076 across three mass windows (m_gb ≈ 0.78—2.4 GeV). Dark energy arises from the residual uncondensed fraction (1 − ε) ∼ 10u207bu2074u2075 of the spin-2 condensate energy. While this value is accommodated within the non-linear saturation framework, its precise first-principles pre—diction remains an open theoretical challenge; it constitutes a well-defined target for future development of the saturation dynamics. We show that the Landau parameter Fu2082 reaches −4.97 ± 1.25 (phenomenological estimate) or −3.65 ± 0.41 (conservative RPA with realistic nuclear forces), both approaching the Pomeranchuk instability thresh—old Fu2082 = −5 with critical density n_c/nu2080 ≈ 1.0—1.2. Finite-temperature enhancement mechanisms at the QCD crossover provide the bridge to the threshold. The framework yields falsifiable predictions including a single nanohertz GW peak (5—15 nHz), distinctive 21-cm power spectrum suppression, and CMB lensing signatures, linking the dark sector to ongoing experiments without introducing new particles or energy scales beyond QC.