Judge Blocks Pentagon Blacklist of Anthropic as Unlawful

A federal judge in California has permanently enjoined the Pentagon from enforcing its designation of Anthropic as a national security supply chain risk, ruling that the move amounted to unlawful retaliation against the artificial intelligence company for its public opposition to certain military applications.

The decision by U.S. District Judge Rita Lin, issued late Thursday, centers on the Trump administration’s handling of the Claude model developer after Anthropic declined to permit its technology for mass domestic surveillance or fully autonomous weapons systems. Lin, appointed by former President Joe Biden, found that the Defense Department’s actions violated the company’s First Amendment rights and due process protections by punishing protected speech under the guise of national security.

“The empty invocation of national security is not a blank check to punish and retaliate against government critics,” Lin wrote in the 59-page order. She determined that the supply chain risk label, which effectively barred defense contractors from working with Anthropic, stemmed primarily from the company’s criticism rather than any substantiated threat. The ruling bars federal agencies from enforcing the designation but does not compel the government to contract with Anthropic or prevent it from favoring other AI providers, provided those decisions follow standard regulatory processes.

The dispute traces back to early 2026, when the Pentagon sought to integrate Claude into classified military systems. Anthropic conditioned its participation on explicit limits against uses it viewed as unreliable or ethically fraught, including weapons that operate without human oversight. When negotiations stalled, the administration cut off direct contracts in March and imposed the broader blacklist. Defense officials maintained the steps were essential to protect sensitive supply chains, though the court found little evidence tying the designation to concrete security risks beyond the company’s policy disagreements.

Anthropic welcomed the outcome, stating it remains committed to collaborating with government partners on AI applications that enhance national security. The company has long positioned itself as more cautious than some competitors on high-stakes deployments, advocating for technical safeguards and human accountability in areas like targeting or intelligence collection.

The case arrives amid accelerating efforts to embed advanced AI across defense operations, from logistics and intelligence analysis to potential combat roles. Proponents argue these tools can improve decision speed and reduce personnel exposure, while critics, including some within the tech sector, warn that current models lack sufficient reliability for lethal autonomy or large-scale monitoring without introducing unacceptable errors or mission creep. Lin’s ruling underscores the legal friction that can arise when private firms impose their own constraints on government use of commercial technology.

The administration has signaled it may appeal, consistent with its broader push to assert executive flexibility in national security procurement. Similar tensions have surfaced in prior disputes over data access and technology export controls, where courts have occasionally reined in expansive claims of authority. For now, the order restores Anthropic’s eligibility for indirect defense work while leaving open the question of how future contracts will navigate the company’s stated boundaries.

Observers note that the decision could influence ongoing debates over AI governance in the military, particularly as Congress considers legislation on autonomous systems and oversight mechanisms. The ruling does not resolve underlying technical debates about model performance but does affirm that companies retain some constitutional space to voice reservations without facing administrative punishment. Both sides are likely to continue litigating the scope of permissible restrictions as AI capabilities evolve.