This working paper presents version 9.9 of a proposed atomic-state theory of gravity. The central claim is that the effective gravitational expression of known matter should not be treated as a uniform function of rest mass alone, but as an environmental and state-dependent response rooted at the atomic level. The theory emphasizes that galaxies, rotation curves, weak lensing, merging clusters, and cosmological observables are not the foundation of the model, but macroscopic consequences of an underlying atomic response. Matter must therefore be distinguished by physical state: cold galactic gas, hot shocked plasma, bound stellar systems, diffuse low-surface-brightness matter, compact red systems, ionized early-universe plasma, neutral atomic gas, and warm-hot intergalactic matter. A key scale result is stated explicitly: the characteristic acceleration scale a_tau = H0 × c / (2π) ≈ 1.082 × 10^-10 m/s^2 is derived from the Hubble constant without any free parameter. This scale is compared with the acceleration scale appearing in galactic dynamics and weak-lensing tests. The document summarizes qualitative and diagnostic tests across independent systems, including SPARC rotation curves, KiDS weak-lensing RAR profiles, Bullet Cluster state-separation tests, and cosmological state-transition diagnostics. The paper does not claim a final proof against dark matter. Instead, it formulates a falsifiable research program: if the same atomic-state response can predict rotation curves, lensing, merging-cluster offsets, and cosmological signatures without individually fitted dark halos, then part of what is currently interpreted as dark matter may represent an unmodeled state-dependent gravitational expression of known matter.