This paper introduces canonical anchoring as a structural mechanism by which AI systems fix semantic resolution to stable origins. Unlike visibility or amplification strategies, canonical anchoring establishes a minimal set of reference surfaces — canonical origins, indexed identifiers, timestamped artefacts, and entity graphs — that make a source structurally unavoidable for meaning resolution. The paper explains why certain sources persist across AI systems once convergence occurs, and why durability follows structure rather than popularity. Complements the Ambiguity Elimination paper by describing the post-convergence mechanism that governs persistence, reuse, and long-term stability. Part of the AI Visibility Architecture (AIVA) framework documentation.