Anthropic is pushing back against the Pentagon in a high-stakes legal battle, arguing it cannot control or manipulate its Claude AI once deployed within military systems.

In a detailed court filing submitted to the U.S. Court of Appeals, the company rejected claims that it poses a “supply chain risk,” calling the designation unjustified and damaging to its reputation and operations.

The dispute stems from a broader conflict over how AI should be used in defense, particularly around autonomous weapons and surveillance.

Anthropic maintains that the Pentagon’s actions amount to retaliation following disagreements over ethical limits on its technology, while the government continues to defend its classification and restrictions.

The case has already seen mixed outcomes, with a separate court previously ordering the removal of stigmatizing labels, even as proceedings continue in Washington ahead of scheduled arguments.

The fallout has had direct commercial consequences, including the loss of a $200 million defense contract, which was subsequently secured by OpenAI.

The case underscores intensifying competition and policy tensions in the AI sector, where national security priorities, corporate ethics, and market dominance are increasingly intertwined.