Palantir link pulls Anthropic into Pentagon “all lawful uses” showdown

A behind-the-scenes dispute between Anthropic and the Pentagon has escalated into talk of punitive action, with Palantir’s role as a key deployment channel sitting at the center of the rupture. Senior administration officials have discussed barring the AI startup from military use, according to reporting cited by Semafor.

They report the conflict’s immediate catalyst was a Palantir–Anthropic check-in after Claude appeared on classified systems during monitoring tied to the January seizure of Venezuela’s Nicolás Maduro. A Palantir executive interpreted Anthropic’s questions about the operation as disapproval of that use and relayed concerns to Defense Department officials, deepening distrust. Anthropic disputes that account, saying it has not raised operational concerns with industry partners beyond technical discussions, while emphasizing its national-security posture and classified-network deployments.

At issue is contract language. The Pentagon is pushing an “all lawful uses” standard, while Anthropic seeks carve-outs limiting certain surveillance and autonomous-weapons applications, reported Axios. The Pentagon has floated labeling Anthropic a “supply chain risk,” a rare designation that could pressure contractors to avoid Claude—raising stakes as Anthropic reportedly nears an IPO.