Anthropic, once seen as the “conscience” of the AI industry for rejecting Pentagon and White House offers involving mass surveillance and autonomous weapons, has lost public goodwill after court documents revealed Project Panama. The project involved purchasing physical books, splicing their spines, scanning the pages, and recycling leftovers, with the goal to “destructively scan all the books in the world.” In February, 2026, the US Department of Defence offered full access to Claude, but Anthropic rejected it, despite having signed a 200 million dollar contract with the Pentagon the previous July. Defense Secretary Pete Hegseth responded by banning Anthropic’s products as a “supply chain risk.” Anthropic sued, and courts disqualified the firm from DoD contracts while allowing other government business. The CFO told a federal judge the affair could amount to billions in lost revenue in 2026. The tech community initially supported Anthropic’s stance. Later, Anthropic released new models, Claude Fable 5 and Mythos 5, but within three days, Commerce shut them down globally due to a jailbreak.