Court Rules Pentagon's Blacklisting of Anthropic Unlawful

A federal judge in California ruled that the Pentagon's blacklisting of Anthropic was unlawful retaliation for public criticism of the administration and permanently enjoined the enforcement of the boycott directive.

Compiled by Sergey KostenchukPublished 2026-08-29Updated 2026-09-18

2026-08-29 Safety OpenAI

🛡️ Court Rules Pentagon's Blacklisting of Anthropic Unlawful

A U.S. court ruled that the Pentagon's blacklisting of Anthropic was unlawful 'retaliation' for public criticism of the administration. The government failed to prove supply chain risk — the boycott is permanently enjoined, but there is no obligation to purchase from Anthropic.

🌍 The first court-tested limit on the Department of Defense's authority over AI vendors: 'blacklists' cannot be used as retaliation for criticism without evidence of risk. Vendor selection remains with the Pentagon, but the industry now has a protection mechanism.

👤 For users: WIRED found 'Persistent Mode' in open-source Codex — the agent works in the background, sets its own tasks, and stores memory about the user; OpenAI confirmed testing.

Source 1: https://www.reuters.com/world/us-judge-rules-pentagon-blacklisting-anthropic-unlawful-2026-08-28/ Source 2: https://www.wired.com/story/openai-is-developing-a-persistent-ai-agent/

Sources