Anthropic AI Iran strikes moved forward despite President Donald Trump’s executive order banning the software from federal use.
Recent reports indicate that United States Central Command (CENTCOM) utilized the Claude AI system, developed by Anthropic, to execute air strikes in Iran. This deployment occurred in the narrow window immediately following Donald Trump’s direct order for federal agencies to cut all ties with the tech firm.
According to details shared by the Wall Street Journal, defense officials leaned on the AI to sift through intelligence data and identify specific targets. The software was also used to run complex battle simulations, giving commanders a digital preview of the mission before boots were on the ground or planes were in the air.
When asked about the specific digital tools used during the ongoing engagement with Iran, CENTCOM officials declined to comment. However, the timing highlights a massive disconnect between the administration’s policy and the tactical reality on the front lines.
The controversy stems from a recent directive where Trump told federal agencies to “immediately cease” any involvement with Anthropic. Trump has labeled the company a national security threat, ordering the Pentagon to treat them as a “supply chain risk.” This designation effectively starts a six-month clock to scrub the software from all government servers.
Despite the current political friction, Anthropic’s Claude AI has a deep history with the U.S. intelligence community.
Through high-level partnerships with Palantir and Amazon Web Services, the model was cleared for classified environments long ago. In fact, it played a quiet but critical role in the January operation in Caracas that led to the capture of Nicolás Maduro.
Anthropic remains part of an elite group of AI labs, including OpenAI and Google, that hold massive Pentagon contracts valued at roughly $200 million. Losing this status would be a major blow, both financially and reputationally.
In response to the sudden ban, Anthropic released a statement calling the “supply chain risk” label an “unprecedented action” usually saved for America’s literal enemies. The company made it clear they plan to fight the order in court, noting that they have been a consistent partner on classified networks since mid-2024 and have no intention of walking away from their military support roles.
