Legal

Anthropic: Court declares Pentagon ban unlawful

3 min read

TL;DR Too Long; Didn’t read

A US federal judge has declared the Pentagon's classification of Anthropic as a national security risk unlawful in a 59-page decision. The court views the measure imposed in February/March as unconstitutional retaliation for Anthropic's criticism of military AI policy. The separate exclusion from Pentagon contracts remains in place for now, with a second proceeding in Washington continuing.

A judge's gavel shatters a red chain bearing the Anthropic logo, with the Pentagon building in the background Image generated with GPT Image 2

Key takeaways

  • Judge Rita Lin declares the Pentagon's classification of Anthropic as a security risk unlawful.
  • The court sees this as unconstitutional retaliation for Anthropic's public criticism of military AI policy.
  • Anthropic had refused to release Claude for autonomous weapons or mass surveillance of US citizens.
  • The exclusion from Pentagon contracts remains in place, with a second proceeding in Washington continuing separately.
  • According to Anthropic, the continued contract exclusion could cost the company billions in revenue.

A US federal judge has declared the Pentagon’s classification of Anthropic as a security risk to national security unlawful. Judge Rita Lin ruled in a 59-page decision that the Department of Defense violated the company’s freedom of speech. A second proceeding to exclude Anthropic from Pentagon contracts is ongoing.

Dispute ignites over rules for combat operations

The conflict dates back to early 2026: after the Pentagon, under Defense Secretary Pete Hegseth, had already threatened the measure in February, it officially classified Anthropic in early March as a “supply chain risk” to national security – a category that had previously only been imposed on foreign companies with alleged connections to Chinese or Russian actors. Anthropic thus became the first US company to receive this label from the department. The trigger was Anthropic’s refusal to release its language model Claude for fully autonomous lethal weapon systems or for mass surveillance of US citizens – a restriction that has been part of the company’s public usage policies for years. Shortly after the classification, the Trump administration ordered federal agencies to phase out Claude and other Anthropic services within six months. Anthropic filed a lawsuit on March 9, calling it an unlawful retaliation campaign. As Al Jazeera reports, the legal dispute followed months of public exchanges between the company and the department. The case is part of a broader power struggle between major AI providers and the US government over control of military AI usage – a dispute in which Anthropic recently surpassed Nvidia in lobbying spending to secure political influence in Washington.

Court accuses department of unconstitutional retaliation

In her decision at the federal court in San Francisco, Judge Lin reaches a clear conclusion: the risk classification violated the First Amendment because it was intended to punish criticism of the Department of Defense, and Anthropic was also denied the hearing required under the Fifth Amendment before the measure took effect. The decision by Secretary Hegseth also violates applicable administrative law and was made without a viable factual basis, making it arbitrary. The reasoning of the judgment states, in essence, that mere reference to national security is not a blank check to punish and retaliate against government critics. A Pentagon spokesperson initially did not comment on the ruling. Observers expect an appeal by the government before the relevant appellate court, though no official timeline has been set. Already in March, Lin had temporarily ruled in favor of Anthropic through a preliminary injunction, labeling the retaliatory intent a “classic” violation of freedom of speech. The current decision now confirms that assessment on the merits.

Contract ban remains in place despite the win

The legal success only partially resolves the situation for Anthropic: the exclusion of the company from Pentagon procurement contracts is the subject of a separate proceeding before a court in Washington, which had already denied an emergency application from Anthropic in April. Even a win in both proceedings would not automatically obligate the department to resume working with the company, as NBC News reports. An Anthropic spokesperson said the company welcomes the ruling and wants to keep working productively with the government so that all Americans benefit from the technology. Anthropic executives put the potential revenue losses from the continued exclusion at a billion-dollar figure – independently unverified. More is at stake for the company than a single contract: just days earlier, investors in a new funding round floated a valuation of two trillion dollars, with prospects for government business factoring into that figure.

What matters now is how the still-open Washington proceeding over the actual contract ban turns out, and whether the government appeals the current ruling. Until then, Anthropic remains locked out of one of the US government’s largest AI customers, while the broader fight over military guardrails for AI models between Washington and Silicon Valley keeps intensifying.

Frequently asked questions

What does the classification as a 'supply chain risk' mean concretely?

The designation is based on a federal procurement security law and allows authorities to remove technology from an affected provider from their own systems. So far, it has almost exclusively affected companies with alleged ties to China or Russia.

Can the Pentagon now work with Anthropic again?

Not automatically. The separate exclusion from procurement contracts is part of a second proceeding ongoing in Washington and remains in place independently of the current ruling.

Why did Anthropic refuse to fully release Claude for military use?

The company fundamentally opposes the use of its models in fully autonomous weapon systems without human control as well as for mass surveillance of US citizens.

Is the Pentagon appealing the ruling?

That is open at this time. The department has not commented on it initially, but observers consider it likely that an appeal will be filed with the relevant appellate court.

What financial consequences does the dispute have for Anthropic?

According to company representatives, the continued contract exclusion threatens revenue losses in the billions. An independent confirmation of this figure has not yet been provided.

Sources (6)
  1. Anthropic PBC v. U.S. Department of War, 3:26-cv-01996 (Court file, CourtListener)
  2. US judge blocks Pentagon blacklisting of AI firm Anthropic – Al Jazeera
  3. Federal judge blocks Pentagon blacklisting of Anthropic, calling it 'illegal and baseless' – NBC News
  4. Judge rules that the Pentagon's Anthropic ban was 'illegal and baseless' – Engadget
  5. Anthropic vs. Pentagon: US government acted 'unlawfully and without basis' – heise online
  6. One Government, Two Voices: Recent Anthropic vs. Pentagon Developments – Jones Walker LLP (Legal analysis)

Your AI update for the work week

Once a week, the most important AI news – plus one practical tip to try right away. No spam, unsubscribe anytime.

← Back to the blog