WASHINGTON, D.C. — The dispute centers on a fundamental question of power: Who controls the “red lines” of AI in combat? Anthropic, led by CEO Dario Amodei, has insisted on contractual guarantees that its flagship model, Claude, will not be used for fully autonomous lethal weapons or mass domestic surveillance. The Pentagon has rejected these limits, asserting that once the military buys a tool, it determines how it is used under the law.
1. The “Supply-Chain Risk” Designation
By labeling Anthropic a “supply-chain risk,” the administration has triggered a massive ripple effect:
- Contractor Ban: Effective immediately, any private contractor (including giants like Palantir or Lockheed Martin) doing business with the military is barred from using Anthropic’s services.
- Six-Month Phase-Out: Federal agencies have been given a 180-day window to migrate away from Claude, which is currently the only AI model embedded in several highly classified military systems.
- Legal Challenge: Anthropic announced late Friday it would challenge the designation in court, arguing that Secretary Hegseth lacks the statutory authority to apply this label to an American company over a policy dispute.
2. OpenAI and xAI: The Ready Alternatives?
As Anthropic is phased out, the Pentagon is already moving toward more “compliant” partners:
- Elon Musk’s xAI: The Pentagon recently signed an agreement to deploy Grok in classified environments. Unlike Anthropic, xAI has reportedly agreed to the “all lawful use” standard without specific ethical carve-outs.
- OpenAI’s Delicate Balance: While the administration has approached Sam Altman to fill the gap, Altman reportedly told employees he shares Anthropic’s “red lines” regarding autonomous killing and surveillance. However, OpenAI is still in active negotiations to expand its footprint in the Department of War.
3. Ideological Warfare
The rhetoric from the White House has turned sharply political. On Truth Social, President Trump branded Anthropic’s leadership as “Left-wing nut jobs” and “radical woke” for trying to “strong-arm” the Department of War. He warned of “major civil and criminal consequences” if the company fails to cooperate during the transition period.
The Great AI Divorce: Current Status
| Feature | Anthropic Position | Pentagon/Trump Position |
| Autonomous Weapons | Strictly prohibited (no “human-out-of-loop”). | Demand “all lawful use” for operational flexibility. |
| Domestic Surveillance | Barred to protect democratic values. | Won’t sign “veto power” over to a private firm. |
| Legal Status | Challenging “Supply-Chain Risk” in court. | Invoking “Full Power of Presidency” for compliance. |
| Replacement | Working on “smooth transition.” | Accelerating Grok (xAI) and Gemini (Google) tests. |