Skip to main content

Judge questions Pentagon’s Anthropic supply-chain-risk designation

6 OCTOBER 2026·2 MIN READ·4 SOURCES

U.S. District Judge Rita Lin questioned whether the Defense Department acted appropriately in designating Anthropic a supply chain risk during Tuesday’s first major hearing in the dispute. She ended the hearing without ruling and said she would decide Anthropic’s preliminary-injunction motion within the next few days.

Judge questions Pentagon’s Anthropic supply-chain-risk designation

Key takeaways · 4

  • 01

    The court has not yet ruled on Anthropic’s preliminary-injunction motion; Lin said she would decide within the next few days.

  • 02

    Anthropic retained contract prohibitions on Claude’s use for lethal autonomous warfare and mass surveillance of Americans.

  • 03

    The Department asked Anthropic for permission to use Claude for all lawful purposes.

  • 04

    The administration’s sabotage argument concerned a possible future software update, and the hearing record cited here did not establish Anthropic’s ability to push one.

Lin challenges the government’s rationale

Lin questioned whether the government had tailored its response to the national-security concern it described.[1] She said the effort to bar Anthropic from government work appeared to be an attempt to cripple the company and punish it for criticizing the government’s contracting position.[1] Lin also questioned whether a company could be designated a supply chain risk simply for insisting on contract terms and asking questions.[1] Her comments indicate skepticism during the hearing, but do not amount to a ruling on the injunction.[1]

The contract dispute centers on permitted uses

Anthropic retained contractual prohibitions on Claude’s use for lethal autonomous warfare and mass surveillance of Americans.[2] The Department asked Anthropic for contractual permission to deploy Claude for all lawful uses.[2] Anthropic was raising concerns to the Department of War about military uses of its technology that were within the usage terms.[1] The dispute therefore involves both the government’s desired contractual flexibility and Anthropic’s limits on certain uses; the hearing evidence does not say that the parties resolved those differences.[2][1]

The government cited a possible software update

The administration argued that Anthropic could pose a national-security risk by sabotaging the military through a future software update.[1] One possibility raised was that Anthropic could attempt to add a “kill switch” if it disagreed with how Claude was being used.[1] However, a speaker could not definitively say whether Anthropic had the power to push such an update.[1] Lin acknowledged that Defense Secretary Pete Hegseth could limit Anthropic’s national-security applications, while questioning whether the Pentagon went too far by designating the company a supply chain risk.[1]

The injunction is still undecided

Lin concluded the hearing without ruling and said she would decide Anthropic’s preliminary-injunction motion within the next few days.[1] The legal backdrop includes a September 25 D.C. Circuit decision upholding the Pentagon’s designation, according to MeriTalk.[3] In a separate related lawsuit, a federal judge ruled against the government, and that ruling remains in effect.[4] These developments concern related litigation, but the evidence provided does not explain how they affect the district court’s pending motion.[1][3][4]

For government buyers and vendors, the dispute highlights how contract limits on AI use can become intertwined with national-security assessments. Professionals should distinguish the government’s stated risk theory from what the court has decided: Lin has not ruled on the pending motion.

Why it matters
Story quiz

Test yourself on this story — 1 question.

Create a free account to take the quiz, earn XP, and get a daily session built for your industry.

Take the quiz

How this developed

  1. 6 October 2026

    Judge questions Pentagon’s Anthropic supply-chain-risk designation

Sources

AI fluency, one session a day, built for your work.