DOD Labels Anthropic a Supply-Chain Risk — Domestic Lab, Adversary Tool

Supply-chain risk designations were built for foreign adversaries. The proprietary shock: the Pentagon turned that tool on a U.S. frontier lab that would not drop Claude bans on fully autonomous weapons and mass domestic surveillance.

U.S. Department of Defense designated Anthropic a supply-chain risk (March 4) after negotiations collapsed over broader access without Anthropic’s acceptable-use limits. Context: reported ~$200 million contract talks; President Trump’s order for agencies to cease using Anthropic technology. Anthropic received a confirming letter March 4. Secretary of War Pete Hegseth announced no military contractors could conduct commercial activity with the company; federal agencies directed to phase out Anthropic products (some reports noted continued use in certain operations despite the order). CEO Dario Amodei had publicly drawn those red lines. Anthropic: action retaliatory and not legally sound; plans to challenge in court (follow-on emails/memo files March 6).

Claims vs checks

Designation and Anthropic response are Anthropic statement + Pentagon/Hegseth announcements, corroborated by The Hill, CNBC, Washington Post. $200M figure is reported contract context. “First U.S. AI company” framing is contemporaneous analysis — verify against prior designations before treating as legal novelty forever.

Limits

• Designation ≠ criminal finding.
• Phase-out compliance can be uneven across agencies.
• Litigation and March 6 removal memo are separate beats.

Sources

• Anthropic: “Where things stand with the Department of War” (March 5, 2026; referencing March 4 letter).
• Pentagon announcements and Secretary Hegseth X posts (March 4, 2026).
• The Hill, CNBC, Washington Post on the designation and contract dispute (early March 2026).
• Anthropic prior statements on acceptable use policy.