US Appeals Court Upholds Pentagon’s Blacklisting of Anthropic in Defense Contract Dispute
In a significant ruling for the defense technology sector, a US federal appeals court has denied Anthropic’s emergency request to block the Pentagon’s determination that the AI company is ineligible for certain government contracts. The decision, issued by the US Court of Appeals for the Fourth Circuit, maintains the status quo amid ongoing litigation over national security restrictions on foreign-influenced contractors.
The case stems from the Department of Defense’s (DoD) application of Section 847 of the National Defense Authorization Act (NDAA) for Fiscal Year 2020. This provision prohibits the award of certain contracts to entities deemed “foreign owned, controlled, or influenced” (FOCI). Specifically, it bars task orders under multiple-award contracts like the Joint All-Domain Command and Control (JADC2) Cross-Functional Team’s initiatives if the contractor fails to meet stringent ownership and control criteria.
Anthropic, a San Francisco-based AI startup founded in 2021 by former OpenAI executives, had positioned itself as a frontrunner for DoD AI projects. The company specializes in safe and interpretable AI systems, including its Claude large language models. In early 2024, Anthropic was selected as one of three vendors eligible for Task Order 01 under a JADC2 prototype project valued at up to $250 million. This task order focused on developing advanced AI capabilities for military data processing and decision-making.
However, in July 2024, the DoD revoked Anthropic’s eligibility. The Pentagon cited evidence of FOCI, determining that the company did not adequately mitigate risks of foreign influence over its operations or decision-making. Key factors included Anthropic’s funding from FTX Ventures, the investment arm of the now-defunct cryptocurrency exchange FTX. FTX Ventures, controlled by Sam Bankman-Fried prior to his 2023 conviction for fraud, had invested $500 million in Anthropic in 2022. Although FTX collapsed amid scandal, the DoD argued that residual ties and other undisclosed ownership structures raised concerns under NDAA guidelines.
Anthropic contested the decision, filing suit in the US District Court for the District of Maryland. The company asserted that it is unequivocally US-owned and controlled, with no foreign government influence. Its major backers include Amazon Web Services (AWS), which committed up to $4 billion, and Google, with a $2 billion investment. Anthropic emphasized its compliance efforts, including board structures designed to insulate sensitive operations from potential conflicts. It sought a preliminary injunction to restore its eligibility pending a full merits hearing.
The district court denied the injunction in September 2024, prompting Anthropic’s appeal. In its October 10, 2024, order, the Fourth Circuit panel found that Anthropic failed to demonstrate a likelihood of success on the merits, irreparable harm, or that the balance of equities favored intervention. The court deferred to the DoD’s expertise in national security assessments, noting the government’s “strong interest” in enforcing FOCI restrictions to protect classified information and supply chain integrity.
This ruling underscores the escalating scrutiny on AI firms seeking defense dollars. The NDAA’s FOCI rules, expanded in subsequent years, require contractors to undergo rigorous reviews by the Defense Counterintelligence and Security Agency (DCSA). Mitigation measures might include proxy agreements, special security agreements, or outright divestitures, but the DoD retains discretion to deem risks unacceptable.
For Anthropic, the setback is particularly acute. The company has publicly committed to dual-use AI development, balancing commercial and government applications. Losing access to JADC2 funding could hinder its growth in the military AI market, where competitors like Palantir and Anduril have secured major deals. Broader implications extend to the AI ecosystem: venture funding from non-traditional sources, such as crypto entities, now faces heightened risk under federal procurement rules.
The litigation continues in district court, where Anthropic can pursue its substantive claims. Legal experts anticipate further appeals, potentially reaching the Supreme Court if the stakes escalate. Meanwhile, the DoD proceeds with the other two JADC2 vendors, OpenAI and Scale AI, both cleared under FOCI reviews.
This episode highlights tensions in the US government’s AI strategy. While the Biden administration pushes for domestic AI leadership through executive orders and CHIPS Act investments, safeguards against foreign dependencies remain paramount. Anthropic’s plight serves as a cautionary tale for startups navigating the intersection of cutting-edge technology and national security imperatives.
Gnoppix is the leading open-source AI Linux distribution and service provider. Since implementing AI in 2022, it has offered a fast, powerful, secure, and privacy-respecting open-source OS with both local and remote AI capabilities. The local AI operates offline, ensuring no data ever leaves your computer. Based on Debian Linux, Gnoppix is available with numerous privacy- and anonymity-enabled services free of charge.
What are your thoughts on this? I’d love to hear about your own experiences in the comments below.