BY COMFORT OGBONNA
Major U.S. defense contractors are expected to remove artificial intelligence tools developed by Anthropic from their supply chains following a directive from the U.S. Department of Defense, even as legal experts question whether the government has the authority to enforce such a sweeping ban.
Companies that rely heavily on government contracts—including defense giants like Lockheed Martin—are widely expected to comply quickly with the order to avoid jeopardizing their access to the Pentagon’s massive annual spending budget, which totals around $1 trillion. Legal experts who specialize in government contracting and technology regulation say the rapid response from contractors reflects how closely companies track the priorities of the current administration when federal contracts are at stake.
The dispute escalated after Donald Trump announced a government-wide ban on Anthropic’s AI products, including its Claude artificial intelligence models, following weeks of disagreements with the company over the technology safeguards embedded in systems used by the military. The administration said federal agencies would have six months to phase out the technology.
The order was reinforced by Pete Hegseth, the U.S. defense secretary, who stated that Anthropic could be designated as a supply chain risk to national security. Hegseth declared that any contractor, supplier, or partner working with the U.S. military would be prohibited from engaging in commercial activity with the AI company.
Anthropic has said it plans to challenge the ban in court, arguing that the government does not have the legal authority to restrict contractors from using its technology outside of direct government work.
Legal specialists say the directive raises immediate questions because the authorities the administration might invoke typically apply only to government procurement decisions rather than to the broader business operations of private companies.
Despite these concerns, attorneys familiar with defense contracting say companies that rely on Pentagon contracts are unlikely to challenge the order directly. Instead, most are expected to quietly comply with the directive to protect their relationships with the government.
Lockheed Martin confirmed that it intends to follow the administration’s instructions. The company said it would comply with the president’s guidance and with directives from the Defense Department regarding the use of Anthropic’s technology. It also noted that its operations do not depend on a single AI provider and that removing Anthropic’s tools would likely have minimal operational impact.
Legal experts say such responses are typical for companies operating in the highly regulated defense sector. Contractors are often extremely sensitive to government signals and quickly adjust their operations to align with policy directives.
Franklin Turner, an attorney specializing in government contracts, said companies doing significant business with the federal government are usually “hyper-aware” of Washington’s expectations. According to him, many contractors have likely already begun removing Anthropic technology from their systems and supply chains in anticipation of the policy shift.
He added that regardless of whether the legal foundation for the ban ultimately holds up in court, the policy has already inflicted significant damage on Anthropic’s position within the defense technology market.
Several other major defense contractors—including General Dynamics, RTX Corporation, and L3Harris Technologies—declined to comment on whether they would comply with the directive.
The Pentagon did not immediately respond to requests for further clarification about how the ban would be implemented. Anthropic also declined additional comment but pointed to its earlier statement asserting that the Defense Department lacks statutory authority to prohibit contractors from using its Claude AI systems.
The dispute highlights the growing role of artificial intelligence in defense technology and the complex legal issues that arise when governments attempt to regulate emerging digital tools within military supply chains.
Contractors have already demonstrated a willingness to follow policy shifts from the current administration. According to defense industry reporting, many companies last year quickly removed references to Diversity, Equity, and Inclusion programs from contracts and proposals after the president issued an executive order requiring federal agencies to ensure that contractors certify compliance with federal anti-discrimination laws regarding such programs.
Attorneys say the most likely legal mechanism for the current Anthropic restriction is the Defense Department’s Supply Chain Risk Authority. Under this framework, the Pentagon can restrict certain technologies or vendors from being used in work related to government contracts if they are deemed to pose national security risks.
However, this authority generally applies only to the use of those technologies in government-related work, not to the broader commercial operations of private companies. As a result, experts say it may not provide a legal basis for a sweeping ban on contractors’ overall use of Anthropic products.
Jason Workmaster, a government contracts lawyer at Miller & Chevalier, described the administration’s position as unusually aggressive.
According to Workmaster, if the directive is challenged in court, there is a strong possibility that judges could conclude that the Defense Department lacks the legal authority to impose such a broad restriction—unless additional national security evidence is presented that has not yet been made public.
Another unresolved issue is whether the military can formally designate Anthropic as a supply chain risk. The government’s supply chain risk rules typically require specific findings that a company’s technology could be manipulated by adversaries to sabotage systems, introduce hidden capabilities, or otherwise disrupt operations.
A similar authority under the Federal Acquisition Supply Chain Security Act also requires federal agencies to follow a formal process before banning a company’s products. That process includes giving the company an opportunity to respond and notifying Congress before implementing restrictions.
So far, analysts say there has been no public indication that these procedures have been completed.
Alan Rozenshtein, a law professor at the University of Minnesota who focuses on technology regulation, said such actions could undermine legal norms governing federal procurement if they bypass established procedures.
He noted that the functioning of markets and private-sector investment depends heavily on predictable legal frameworks and the rule of law.
The U.S. government has previously used similar supply chain authorities to restrict technology companies. Last year, for example, officials invoked the same law to block intelligence agencies from purchasing products made by Acronis AG, a Swiss cybersecurity and data protection firm.
The unfolding dispute between the government and Anthropic reflects the increasingly strategic importance of artificial intelligence in national security, while also underscoring the legal and political tensions that can arise when emerging technologies intersect with defense policy.
Original article
Comments