Pentagon formally designates Anthropic a supply chain risk amid feud over AI guardrails

The U.S. military has taken a significant step by designating artificial intelligence firm Anthropic as a supply chain risk, a move that could potentially cut off the company from military contracts. This decision comes after a standoff between the Trump administration and Anthropic regarding the use of its Claude model in military operations.

Anthropic, the only AI company deployed on the Pentagon’s classified networks, has been pushing for guardrails that would explicitly prohibit the U.S. military from using its AI technology for mass surveillance on Americans or for fully autonomous weapons. On the other hand, the Pentagon argues that it needs the flexibility to use Anthropic’s Claude model for “all lawful purposes” and asserts that current regulations already prohibit the unauthorized uses of AI.

Defense Secretary Pete Hegseth recently announced that Anthropic would be phased out from government contracts over a period of six months. However, the company had not received formal notification of this decision until recently. The military has been using Anthropic’s Claude model in recent strikes on Iran, although the exact nature of its deployment remains unclear.

CEO Dario Amodei has been in talks with the Pentagon to address the situation and has emphasized the importance of setting boundaries to protect American values. He believes that AI technology could potentially grant the government excessive surveillance powers that go against fundamental principles. Amodei has insisted on maintaining guardrails to prevent the misuse of AI in sensitive areas such as mass surveillance and autonomous weapons.

The Pentagon has offered compromises to address Anthropic’s concerns, but the company has deemed them inadequate. Tensions escalated last week, with Trump administration officials accusing Anthropic of imposing its values on the military and hindering its operations. President Trump ordered federal agencies to cease using Anthropic’s technology, with the Defense Department given a six-month timeframe to phase it out.

In response to these developments, Anthropic’s rival, OpenAI, announced a new agreement with the military. A senior Pentagon official emphasized the importance of allowing the military to use technology for lawful purposes without external restrictions. Amodei criticized the Trump administration’s actions as retaliatory and punitive, reaffirming his commitment to upholding American values and supporting national security.

Despite the disagreements, Amodei remains steadfast in his belief that challenging the government is a fundamental right and a patriotic act. He asserts that Anthropic has always prioritized the values of the country and will continue to do so in the face of adversity.