This year has seen a shift in Washington’s treatment of artificial intelligence, beginning in February when Defense Secretary Pete Hegseth declared the leading AI firm Anthropic a “supply chain risk.” In April, Anthropic announced its Mythos model would be released to only a handful of organizations due to extraordinary abilities to break into computer systems. In June, Senator Mark Warner told the Senate Banking Committee that the tool “broke into almost all of our classified systems, not in weeks, but in hours.” The Trump administration initially opposed AI regulation, advised by tech figures like David Sacks and Marc Andreessen, but the national security community has since recognized that the latest AI models pose a clear and present danger. The Commerce Department has now given Anthropic the green light to give its Mythos 5 models to “trusted partners,” and OpenAI announced its newest model would be restricted to a few government-approved organizations.