Federal Court Rules Government’s Anthropic Supply Chain Designation Was Unlawful Retaliation: What Government Contractors Need To Know

This is the third installment in our coverage of the Anthropic/Department of War dispute. Our first alert addressed the Trump administration’s February 27 directives barring federal contractors from using Anthropic and its Claude platform, including Secretary Hegseth’s designation of Anthropic as a supply chain risk to national security and the resulting requirement that all DoW contractors sever commercial ties with the company. Our second alert covered the March 26 preliminary injunction, in which Judge Lin found Anthropic had demonstrated a likelihood of success on the merits and was suffering irreparable harm from the challenged actions. This post covers where things stand now following the latest Federal Court ruling.

The Ruling

On August 27, 2026, U.S. District Judge Rita F. Lin of the Northern District of California granted summary judgment largely in favor of Anthropic PBC, ruling that the DoW’s designation of the company as a supply chain risk to national security was “illegal and baseless.” In a 59-page order, Judge Lin concluded that the designation was not based on any credible threat, but was instead retaliation for Anthropic’s public criticism of the Trump administration’s position during contract negotiations over the use of its AI model, Claude.

As a reminder: this dispute traces back to a breakdown in contract negotiations between Anthropic and the Pentagon. Anthropic maintained contractual restrictions on the use of Claude for two categories of application it deemed unsafe: mass domestic surveillance and fully autonomous lethal weapons systems. The Pentagon insisted that once it acquires a technology, it must be free to use the tool for any lawful purpose, and that it would not accept private vendors dictating operational use cases. The government’s response went well beyond choosing a different AI provider. The President ordered all federal agencies to cease using Anthropic’s products, and Secretary of War Hegseth designated Anthropic a supply chain risk under 10 U.S.C. § 3252, a label historically reserved for foreign adversaries. The designation required all military contractors, suppliers, and partners to sever commercial ties with Anthropic, threatening the company’s up-to-$200 million Pentagon contract and its broader commercial relationships with defense-adjacent businesses.

Key Legal Findings

The court’s order addressed five claims. Anthropic prevailed on four of them.

1. First Amendment Retaliation. The court found that Anthropic’s public statements defending its right to restrict the use of Claude for mass surveillance of Americans and lethal autonomous warfare constituted protected speech on a matter of great public importance. The undisputed record showed that this speech was a “substantial motivating factor” behind the government’s decision to designate Anthropic a supply chain risk and bar it from all federal defense work. Both Secretary of War Pete Hegseth and President Trump made public comments explicitly tying the designation to Anthropic’s speech. The court also pointed to the DoW’s own internal memorandum, which identified Anthropic’s engagement with the press as the basis upon which its risk level was escalated. The court rejected the government’s argument that speech with any connection to contract negotiations cannot qualify as protected speech under the First Amendment.
2. Due Process. Anthropic had no notice or opportunity to challenge the designation before the government publicly barred it from all federal work and blacklisted it with defense contractors. The court further noted that the DoW’s risk assessment was “based on a core misunderstanding about how Anthropic’s technology works,” a misunderstanding the company could have “readily dispelled” if given the chance to respond.
3. Administrative Procedure Act. The court found the supply chain designation was “in excess of statutory authority, contrary to law, and arbitrary and capricious.” The government conceded that Secretary Hegseth lacked the authority to prohibit contractors doing business with the military from conducting any commercial activity with Anthropic. Nothing in the administrative record showed that Anthropic’s conduct met the statutory requirements for a “supply chain risk” under 10 U.S.C. § 3252. Which was designed to address covert acts of sabotage and subversion of national security systems.
   
   Judge Lin highlighted the implausibility of the government’s reasoning. DOD pointed to Anthropic’s “hostile public statements” and its insistence on usage restrictions as evidence of a sabotage risk, but never explained the “inferential leap” from those positions to a belief that Anthropic would “poison” its own models. The court noted that the administrative record supporting the designation was generated almost entirely within a two-day window. As Judge Lin observed, “sabotage would ordinarily be a surprising culmination to months of ‘cordial’ negotiations.”
4. Evidence of Pretext. The court’s most pointed observations may have been about the government’s own behavior. Immediately after designating Anthropic a supply chain risk, the DoW continued to pursue a contract with the company. The day after the designation was finalized, Under Secretary of War Emil Michael, who authored the sole analytical memorandum underlying the supply chain designation, emailed Anthropic’s CEO Dario Amodei stating that the parties were, in his words, “very close here” on contract terms. The government has also been discussing collaboration with Anthropic on its newer “Mythos” model across an array of sensitive contexts, including defensive cybersecurity. Judge Lin’s conclusion was direct: “None of that is consistent with a genuine fear that Anthropic is a saboteur who would poison its software to harm national security.”
5. Ultra Vires Separation of Powers. Anthropic also argued that the Presidential Directive exceeded the President’s constitutional and statutory authority. The court rejected this claim, finding that Anthropic’s passing references to procurement statutes and Congress’s power of the purse were insufficient to show that the President clearly exceeded his background constitutional authority. This was the only claim on which the court ruled against Anthropic.

What This Means for Government Contractors

So, what does this mean in practice for companies that hold (or are pursuing) federal contracts, particularly those integrating AI products into their operations?

The Designation Is No Longer in Effect

The court’s summary judgment order follows the preliminary injunction issued in March and formally removes the legal basis for the DoW’s supply chain risk designation of Anthropic. Government contractors and their suppliers are no longer required to sever commercial ties with the company as a condition of doing business with the DoW.

For many contractors, the original directive triggered real operational consequences: certifications of non-use of Claude, refactoring of internal tooling and pipelines to alternative models, and unwinding of integrations that had been embedded across proposal drafting, code generation, document review, and analytics workflows. This ruling means those emergency measures are no longer legally required. Companies that made those transitions should evaluate whether to resume their Anthropic relationships, taking into account the current state of their alternative deployments, the costs and benefits of switching back, and any contractual commitments made to replacement vendors in the interim.

AI Procurement Decisions Should Be Revisited

The designation created significant uncertainty for organizations that relied on, or were evaluating, Anthropic’s products for commercial or government-adjacent work. That uncertainty rippled beyond direct DoW contractors. The GSA pulled Anthropic from USAI.gov and the Multiple Award Schedule, and civilian agencies followed the Pentagon’s lead in unwinding Anthropic-based solutions. Competitors moved quickly to fill the gap, with rival AI providers positioning their models for defense and classified use.

With the designation struck down, the competitive landscape has shifted again. Companies should revisit AI procurement decisions that were deferred or altered because of the designation. This includes decisions to adopt alternative models that may have been driven more by regulatory risk than by the merits of the technology. It also means reassessing vendor relationships entered into under time pressure during the six-month wind-down period. The ruling does not require any contractor to return to Anthropic, but it does restore the company as a viable option in the federal AI ecosystem.

The Government’s Supply Chain Designation Authority Has Limits

One of the most important aspects of this ruling is what it says about the scope of 10 U.S.C. § 3252. The court concluded that Section 3252 was designed to address covert acts of sabotage and subversion of national security systems, not to penalize vendors who take public positions the government disfavors during commercial negotiations. The government could not point to any technological capabilities unique to Anthropic that drove its risk assessment, and the administrative record supporting the designation was generated almost entirely within a two-day window.

 For contractors and AI vendors alike, the ruling establishes a meaningful precedent: the government must demonstrate a genuine, articulable security basis for a designation, not simply invoke national security as a pretext. Companies that face similar designations in the future now have a concrete legal framework for challenging them.

Due Process Protections Apply to AI Vendors

The court’s due process finding carries independent significance. Anthropic had no notice or opportunity to respond before the government publicly barred it from all federal work and blacklisted it with defense contractors. The court emphasized that the DoW’s risk assessment was based on a fundamental misunderstanding of how Anthropic’s technology works, one that the company could have readily corrected if given the chance.

This matters for the AI and technology sectors broadly. AI products are technically complex, and government decision-makers may not always understand how a model’s architecture, training data, or usage restrictions actually function. The court’s ruling reinforces that companies are entitled to a meaningful opportunity to address factual errors before a designation takes effect. That protection is especially valuable in an environment where AI procurement decisions are being made at speed and where the stakes of a supply chain designation extend well beyond a single contract.

AI Governance Frameworks Should Be Updated

Supply chain risk designations were historically confined to hardware, telecommunications, and companies owned by U.S. adversaries. This case demonstrates that the government is willing to extend those tools into the cloud and AI stack, targeting a domestic company over a policy disagreement rather than a traditional national security threat. Even though this particular designation was struck down, the underlying regulatory mechanism remains in place.

Organizations should update their internal AI policies and vendor governance frameworks to account for the possibility that AI providers could be subject to future supply chain designations or similar government actions. That means building flexibility into AI procurement strategies, maintaining contingency plans for vendor disruption, and ensuring that AI governance playbooks address not only technical and privacy risks but also the regulatory and political risks that can arise when AI companies take public positions on sensitive policy questions.

A Potential Appeal Remains on the Table

The ruling is a decisive win for Anthropic at the district court level, but it may not be the final word. The government may seek to appeal to the Ninth Circuit, as it did at the preliminary injunction stage, although that earlier appeal was stayed by mutual agreement pending a related D.C. Circuit case and was never decided on the merits. Companies should continue to monitor developments in this matter, and be prepared to adjust their compliance posture if the appellate landscape shifts. For now, the summary judgment order stands, and the designation is invalid.

The Bottom Line

Government contractors and their suppliers are no longer required to sever commercial ties with Anthropic company as a condition of doing business with the DoW.

Moreover, this is a significant ruling for the broader question of how far the federal government can go in using supply chain designations against domestic AI companies. And it may not be the last word. We will continue to track the case as it develops.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.