Pentagon Terminates Anthropic Contract Following Dispute Over AI Safety Guardrails
Washington, Sunday, 1 March 2026.
Defense Secretary Pete Hegseth has terminated Anthropic’s $200 million contract after the firm refused to lift safeguards against mass domestic surveillance and autonomous weapons. In a significant escalation, the Pentagon has now designated the safety-focused AI lab a “supply chain risk,” ordering an immediate federal phase-out of its technology.
Escalation to Termination
Following our previous report on Anthropic’s refusal to accept the Pentagon’s ultimatum regarding ethical safeguards [1], the standoff has culminated in a decisive severance of ties between the federal government and the San Francisco-based AI firm. On Friday, February 27, 2026, Defense Secretary Pete Hegseth officially terminated Anthropic’s work with the Department of Defense, citing supply chain concerns [2]. This action was immediately compounded by President Donald Trump, who directed all federal agencies to cease using Anthropic’s technology effectively immediately, declaring that the administration would not do business with the company again [4]. While the Pentagon has allotted a six-month grace period to phase out the technology from its systems, the General Services Administration (GSA) moved swiftly to suspend access to Anthropic on the USAi platform and remove it from primary procurement systems as of Friday [2][4].
The ‘Supply Chain Risk’ Designation
The administration’s response extends beyond a simple contract cancellation; the Pentagon has designated Anthropic a “supply chain risk” [2]. This classification is particularly damaging as it not only ends the company’s contract—valued at up to $200 million—but also prohibits other defense contractors from utilizing Anthropic’s technology, effectively blacklisting the firm from the defense industrial base [2][4]. In response to what analysts have termed the most aggressive AI regulatory move by any government to date, Anthropic has vowed to sue Secretary Hegseth over the designation [2][4]. Despite the loss of government revenue, the company projects 14 billion in total revenue for the year, suggesting its commercial business remains robust despite the federal lockout [2].
Competitors Fill the Void
As Anthropic exits the federal stage, its primary competitors are moving quickly to absorb the available defense demand. On Friday, February 27, OpenAI CEO Sam Altman announced a deal to supply AI models to the Pentagon’s classified military networks [2][3]. This agreement follows xAI, led by Elon Musk, reaching a similar agreement earlier in the week regarding classified systems [5]. Unlike Anthropic, OpenAI and xAI have aligned their operational protocols with military requirements, although OpenAI maintains that its technology will still exclude unlawful uses such as domestic surveillance and autonomous offensive weapons [4]. The market responded favorably to OpenAI’s positioning, with the company raising an additional $110 billion on Friday, pushing its valuation to 730 billion [3].
The Core Ethical Dispute
The schism centers on specific “red lines” regarding how artificial intelligence is deployed in combat and intelligence operations. Anthropic CEO Dario Amodei refused to concede to the Pentagon’s demand for “all lawful uses” of the Claude model, specifically citing concerns that the technology could be used for mass domestic surveillance or fully autonomous weaponry without human oversight [4][5]. Amodei argued that while AI is essential for defense, current large language models are not reliable enough to be removed from human control in lethal scenarios [6]. Conversely, the Trump administration and Defense officials view these ethical constraints as a threat to national security, with Undersecretary for Defense Emil Michael accusing Amodei of attempting to personally control military policy and putting the nation at risk [3]. Despite the administrative backlash, Anthropic has received support from within the industry, with over 500 employees from rival firms Google and OpenAI signing an open letter backing the company’s stance [4].