Pentagon's Supply Chain Ban on Anthropic Struck Down by Federal Court

A 59-Page Rebuke of National Security Overreach

US District Judge Rita Lin delivered a decisive blow to the Pentagon's attempt to blacklist Anthropic, ruling that the Defense Department's supply chain risk designation was both illegal and without factual foundation. In a 59-page decision, Lin barred federal agencies from enforcing the restrictions placed on the AI company and its Claude model, finding that invoking national security concerns does not provide unlimited authority to penalize critics of government policy.

The ruling represents a rare judicial check on executive branch power in the AI policy arena, where national security arguments have increasingly been deployed to shape technology decisions. Lin's opinion made clear that such invocations require substantive justification. "The empty invocation of national security is not a blank check to punish and retaliate against government critics," she wrote, adding that neither constitutional principles nor federal statutes permit sweeping penalties based primarily on a company's disagreement with administration positions.

At DailyTechWire, we've tracked how supply chain risk designations have evolved from tools targeting foreign adversaries into mechanisms that can touch domestic firms. This case illustrates the tension between security prerogatives and the boundaries of regulatory action when applied to US-based AI developers.

The February Standoff Over Model Safeguards

The conflict erupted in late February when the Department of Defense, under Secretary Pete Hegseth, pressed Anthropic to strip certain safety features from its AI systems. The specific safeguards at issue have not been publicly detailed, but the pressure came amid broader debates over how tightly AI models should constrain outputs in government and defense contexts.

Anthropic CEO Dario Amodei drew a firm line, stating the company would not permit its model to be deployed for mass surveillance programs or the development of autonomous weapons systems. That refusal set the stage for escalation. Within days, the administration ordered all federal agencies to cease using Claude and other Anthropic services, with a six-month transition period announced for agencies to migrate to alternative platforms.

Negotiations between Anthropic and the Pentagon broke down, and the company became the first US-based firm to receive a supply chain risk designation in this category. Anthropic responded by filing suit in two separate federal courts, challenging both the designation itself and the processes used to impose it.

What the Ruling Does and Doesn't Resolve

Judge Lin's decision addresses one of those two legal challenges, handing Anthropic a victory on procedural and constitutional grounds. The court found that the Pentagon failed to demonstrate a legitimate basis for the designation and that the action appeared retaliatory in nature, tied to the company's public stance on AI safety rather than verifiable security concerns.

However, Anthropic remains under a separate supply chain risk designation pending the outcome of the second lawsuit. That means federal agencies are still technically barred from procuring or deploying Claude until the legal picture fully clears. Even if Anthropic prevails in both cases, the Pentagon retains discretion over whether to resume collaboration. The ruling compels the government to lift unlawful restrictions but does not mandate renewed contracts or partnerships.

The practical impact in the near term may be limited. Many agencies had already begun shifting workloads to competing models from OpenAI, Google, and other providers during the six-month wind-down period. Reversing those migrations will require new procurement decisions, and some agencies may opt to stay with their current vendors to avoid further disruption.

Business Momentum Unaffected by the Ban

Despite the legal and policy turbulence, Anthropic's commercial trajectory has not faltered. The company reported revenue of $11.5 billion in its most recent quarter, a fourteen-fold increase compared to the same period a year earlier. That growth reflects surging enterprise adoption, particularly among financial services, healthcare, and technology firms that have integrated Claude into customer service, document analysis, and internal workflow automation.

Anthropic is widely expected to pursue an initial public offering later this year, with analysts projecting a valuation that could rank among the largest tech debuts on record. The supply chain designation, while politically and operationally disruptive, has not diminished investor appetite or enterprise demand for the company's models.

The disconnect between government restrictions and private-sector momentum underscores a broader pattern in AI policy. Regulatory actions that target specific vendors often fail to slow commercial adoption when customers perceive technical or performance advantages. In Anthropic's case, the company's reputation for thoughtful safety engineering and transparent documentation has resonated with enterprise buyers wary of reputational risk.

Implications for AI Governance and Executive Authority

The ruling arrives at a moment when governments across Asia and the West are grappling with how to regulate AI without stifling innovation or overstepping legal boundaries. Judge Lin's decision is likely to influence how other courts and policymakers assess the use of national security designations in the technology sector.

In South Korea and Japan, regulators have watched the US case closely as they draft their own frameworks for managing AI supply chains and dual-use technology. Singapore's Infocomm Media Development Authority has emphasized transparency and due process in its approach to AI vendor assessments, a posture that aligns more closely with the principles Lin articulated than with the Pentagon's initial action.

The case also raises questions about the durability of AI safety commitments when they conflict with government demands. Amodei's refusal to remove safeguards was framed as a matter of principle, but it carried significant business risk. The fact that Anthropic weathered the designation without material commercial damage may embolden other AI developers to resist pressure to compromise safety features, particularly in jurisdictions with stronger rule-of-law protections.

Looking ahead, the second lawsuit remains unresolved, and the Pentagon could appeal Lin's ruling. But the legal precedent is now established: invoking national security requires more than assertion. It demands evidence, process, and a credible connection between the restriction and a genuine threat. For an industry built on data and logic, that standard feels overdue.