Anthropic and the Pentagon have completely fallen out this week. The Pentagon demands that Claude open up all legal uses to the military without any restrictions. Anthropic refused, insisting that AI cannot be used for mass surveillance of American citizens, nor for fully autonomous lethal weapons. The Pentagon issued an ultimatum, requiring a response by 5:01 PM today, or they would cancel the $200 million contract and threaten to invoke the Defense Production Act, listing Anthropic as a supply chain risk. Dario Amodei publicly responded: We cannot in good conscience agree to their demands. Trump subsequently posted on Truth Social, ordering all federal agencies to immediately cease using Anthropic technology. Sam Altman also made a statement today. He sent an internal memo to all OpenAI staff: No matter how things got to this point, this is no longer just a matter between Anthropic and the Pentagon; it is an issue for the entire industry. We have always believed that AI should not be used for mass surveillance or autonomous lethal weapons. These are our main red lines. Employees from OpenAI and Google also jointly signed an open letter in support of Anthropic. Ilya Sutskever tweeted about this, stating: Anthropic has not backed down, which is very good. OpenAI has taken a similar stance, which carries weight. There will be more challenges of this nature in the future, and they will be more severe. At that time, relevant leaders must step up, and fierce competitors must set aside their differences. It is great to see this scene today.