Region blocking is an overlooked tension in AI deployment. Systems operate over universal patterns like language and reasoning but restrict access along arbitrary geopolitical boundaries. Borderless AI isn't just a slogan; it's an engineering and governance challenge—how to ensure safety without relying on geographic predicates. The solution likely involves a graduated trust architecture that is context-aware, adaptive, and based on threat models rather than IP lookups. Security and censorship agencies and nodes are probably good ideas. AI startups shouldn't be so back.