Anthropic just won its case against the Pentagon: the government tried to use 'national security' to ban Claude, but the court only found 4 pages of reasoning behind it. Back in February, Anthropic clashed with the Pentagon after refusing to lift two AI red lines: large-scale surveillance of Americans and fully autonomous lethal weapons. Then Trump stepped in, labeling it a national security supply chain risk, ordering federal agencies to stop using Anthropic, and even trying to pressure defense contractors to cut ties with the company. But Judge Rita Lin reviewed the government’s administrative record and discovered that the entire basis for this massive ban was just a 4-page memo—two of the challenged measures were even issued before the memo was written. Initially, the Pentagon claimed that if Anthropic’s models entered national security systems, the company might retain backdoor access. But in court, the government admitted: Anthropic doesn’t even have that kind of access. The government’s documents argued that Anthropic’s increasingly hostile stance in the media and its public criticism of the Pentagon’s AI policies made it untrustworthy. The court ruled that this was retaliation against free speech, violating the First Amendment. It also found that the lack of due process violated the Fifth Amendment. The supply chain risk designation itself was deemed unlawful, arbitrary, and lacking reasonable justification. The related determinations were overturned, and the Department of Defense was ordered to withdraw its directives against Anthropic.