Judge Rules Anthropic Blacklist Illegal

A federal judge ruled the Pentagon's Anthropic blacklist illegal retaliation over AI safety limits.

@sharedot

Judge Rules Anthropic Blacklist Illegal

By @sharedot · · 8 pages

  • AI
  • Anthropic
  • Pentagon
  • First Amendment

A federal judge ruled the Pentagon's Anthropic blacklist illegal retaliation over AI safety limits.

Judge voids Pentagon's Anthropic blacklist

U.S. District Judge Rita Lin ruled Thursday that the Pentagon's designation of Anthropic as a national security supply chain risk was illegal, handing the maker of Claude a major court victory.

Her 59-page order found the administration's actions "constituted unlawful retaliation in violation of the First Amendment" and denied Anthropic required due process.

Lin vacated the designation, barred the agencies named in the lawsuit from enforcing President Trump's directive to stop using Anthropic's tools, and ordered the government to rescind related guidance. A March pause on the blacklist is now permanent.

How the standoff began

According to SiliconANGLE, the dispute began in fall 2025, when the Defense Department sought to bring Claude to an internal platform called genai.mil but objected to Anthropic's usage limits.

Anthropic refused to let its models power fully autonomous lethal weapons or mass surveillance of Americans; the Pentagon demanded language permitting "all lawful uses."

After talks collapsed, Defense Secretary Pete Hegseth — who accused the company of "arrogance and betrayal" — designated Anthropic a supply chain risk, the first U.S. company ever publicly labeled one, and Trump ordered federal agencies to stop using its products.

'Blank check' rejected by court

The ruling is striking because the government's core security claim collapsed: officials conceded Anthropic has no backdoor access to deployed systems and its technology is "no riskier to the national security than any other 'black box' artificial intelligence model."

Lin wrote that "the empty invocation of national security is not a blank check to punish and retaliate against government critics," calling the measures "illegal and baseless" and driven by a desire to "make a public example" of Anthropic.

Ars Technica notes she read the supply-chain statute as targeting covert sabotage, not public contract disputes.

A case built on shifting claims

Straight Arrow reports the Pentagon's rationale changed repeatedly during litigation.

An internal memo by Under Secretary Emil Michael warned Anthropic could "disable" or "alter" Claude mid-operation, but his later sworn declaration dropped that claim, and the government pivoted to arguing a hidden flaw could surface in a future model.

Anthropic countered that it cannot access or shut down deployed models and that replacements must be trained from scratch and independently tested by the Defense Department. Lin found no evidence of real-time disruption risk and described a record of "shifting justifications."

Stakes: contracts and a $2T IPO

The designation had barred defense contractors from doing business with Anthropic, and the company warned the blacklist could cost it billions in revenue.

Straight Arrow reports a two-year contract worth up to $200 million with the Pentagon's Chief Digital and Artificial Intelligence Office was hanging in the balance.

The win lands ahead of a planned IPO that The Guardian and Reuters report could value Anthropic at $2 trillion and raise $100 billion, topping SpaceX's June record. CNBC reports the ruling clears a significant hurdle for that listing.

What comes next

The legal fight is not over. A parallel case remains pending before the U.S. Court of Appeals for the D.C.

Circuit, and CNBC reports Anthropic technically remains a supply chain risk until that case is resolved. The Guardian reports the panel includes two Trump appointees skeptical of Anthropic's arguments, and the government may appeal Lin's ruling.

Her order does not force the Pentagon to buy Claude or restart joint work. Anthropic said it remains "focused on working productively with the government to harness AI for our national security."

Sources

  1. cnbc.com › Judge blocks Pentagon blacklist of Anthropic as supply chain risk
  2. theguardian.com › Pentagon's blacklisting of Anthropic was unlawful, US judge rules
  3. arstechnica.com › Trump blacklisting of "woke" Anthropic deemed illegal by federal judge
  4. siliconangle.com › Court rules Pentagon can't ban Anthropic's AI models
  5. san.com › How the government's case for blacklisting Anthropic fell apart

More on AI