Reports have emerged that the National Security Agency is already utilizing Anthropic’s next-generation “Mythos Preview” model, even as the Department of Defense designates the AI company a “supply chain risk to U.S. national security.” This is the pinnacle of Cognitive Enclosure: the state is hoarding the most advanced autonomous hacking tools while publicly banning their use to justify a “security auditing supercycle” that will cost private enterprises billions. President Trump’s order to cease using Anthropic’s AI has effectively created a two-tiered digital reality where the intelligence community operates with tools that are legally off-limits to the public. Anthropic has vowed to fight the designation in court, but the damage to the digital commons is done. The NSA’s reliance on Mythos—a model capable of zero-day exploitation—reveals that the government’s concern isn’t safety, but monopoly. By labeling the model a risk, they ensure that only the “vetted” circle of Project Glasswing partners can access the perimeter of digital defense. For the rest of the world, the “supply chain risk” is a permanent condition, a built-in vulnerability that keeps the state’s autonomous agents at the top of the food chain. The digital perimeter is being enclosed, and the gatekeepers are the only ones with the keys.