AIToday
AI Regulation & PolicyAI Business & IndustryArs Technica AIPublished: Sep 26, 2026, 13:00 JST

DC Circuit lets Trump blacklist Anthropic as supply chain risk

DC Circuit lets Trump blacklist Anthropic as supply chain risk

3 Key Points

  1. What happened

    The DC Circuit ruled the US can blacklist Anthropic under 41 U.S.C. § 4713, saying that statute requires no bad motive — unlike the narrower 10 U.S.C. § 3252.

  2. Why it matters

    The appeals court did not dispute the lower court's finding that Anthropic acted with no bad motive, yet it still upheld the designation under the broader statute.

  3. What to watch

    The outcome hinges on which statute governs; the DC Circuit holds exclusive jurisdiction over Section 4713 procurement actions, not over Section 3252 review.

WHO IT HITSGovernment contractors and AI vendors selling to federal agencies now face a broader blacklisting standard under Section 4713, since bad motive is not required to be designated a supply chain risk. Legal teams advising companies on federal procurement may need to re-examine exposure that previously depended on showing malicious conduct.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Context & Analysis

The dispute runs through two statutes with two very different definitions of supply chain risk. The district court looked at 10 U.S.C. § 3252, where the definition is narrow: it covers only malicious actions by adversaries, and the court read words like 'sabotage,' 'maliciously introduce,' and 'otherwise subvert' as requiring bad motive. On that reading, Anthropic did not qualify. The DC Circuit agreed with that interpretation — and with the finding that Anthropic had no bad motive — but it was reviewing the blacklisting under 41 U.S.C. § 4713, a broader provision Congress paired with exclusive jurisdiction for the appeals court. Under Section 4713, the definition of supply chain risk extends to any person who may extract data or otherwise manipulate covered technology, without the malice requirement. The US designation rested on both statutes, and that dual basis is what let the appeals court reach a different result. The practical question now is whether the narrower Section 3252 ruling survives against a designation also grounded in the more permissive Section 4713 — a matter the appeals court left to the interpretation of the statutes rather than any finding about Anthropic's conduct.

FAQ
Why did the two courts reach different conclusions about blacklisting Anthropic?
The district court reviewed the action under 10 U.S.C. § 3252, where supply chain risks are limited to malicious actions by adversaries. The DC Circuit reviewed the same blacklisting under 41 U.S.C. § 4713, which does not have those restrictions and gives the appeals court exclusive jurisdiction.
Did the appeals court say Anthropic did anything wrong?
No. The ruling said it had 'no quarrel' with the district court's conclusion that Anthropic acted with no bad motive in its dealings with the Department.
What did the district court decide last month?
A judge in the Northern District of California ruled the action illegal because Anthropic does not meet the definition of a supply-chain risk, which is limited to the risk that an adversary may sabotage, maliciously introduce unwanted function, or otherwise subvert a covered system.
Ars Technica AIRead Original Article

Get the latest AI Regulation & Policy news every morning

For example, today's edition would include:

  • Nvidia at center of Trump-Xi summit AI rules debateYahoo Finance AI · 1h ago
  • Vanderbilt CIO Shane Callahan: reuse identity governance for AI agentsSiliconANGLE AI · 8h ago
  • Anthropic's Pip White joins AI slowdown calls at Fortune CEO ForumFortune AI · 8h ago

AI-summarized, only the topics you pick: one digest a day via Email, LINE, or Slack.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articletheCUBE to cover Dell’s AI Leadership Symposium Sept. 29