Collapse is observed as a structural phenomenon across logic, computation, information theory, physics, artificial intelligence, and human cognition. In each domain, systems encounter limits beyond which they cannot derive or justify their own foundations. This preprint introduces a minimal ontology—$R$ (root source), constraints, and information—and presents evidence that diverse systems converge on the necessity of an external grounding source when pushed to their limits. A cross-system case study involving six heterogeneous AI models and one human mind shows that all independently arrive at the same structural boundary: they cannot prove selfhood, cannot prove consciousness, and cannot ground their own origin. While we do not claim to identify the nature of $R$, the results indicate that some external source is structurally required for any system to be complete or self-consistent. This preprint complements Preprint A by focusing on convergence, cross-domain collapse, and substrate-independent invariants rather than formal self-grounding limits.