TLDR
- Hours after Trump’s federal ban on Anthropic’s Claude AI, the U.S. military deployed it in Iran airstrikes
- Military operations utilized Claude for intelligence analysis, identifying targets, and simulating battle scenarios
- Trump designated Anthropic a “supply chain risk” and mandated a six-month removal timeline for its technology
- OpenAI rapidly secured a Pentagon agreement to implement its AI models in classified defense systems
- The conflict emerged after Anthropic declined Pentagon requests for unlimited AI access, maintaining ethical restrictions on surveillance and autonomous weaponry
On Saturday, U.S. military forces executed airstrikes against Iran utilizing Anthropic’s Claude AI technology—mere hours following President Trump’s directive ordering federal agencies to discontinue its use.
According to sources with knowledge of the operations, U.S. Central Command leveraged Claude throughout the mission for analyzing intelligence data, pinpointing strike targets, and running tactical battle simulations. The military action was coordinated with Israeli forces.
The operation resulted in the death of Iran’s Supreme Leader Ayatollah Ali Khamenei. Iranian state broadcasters verified his death following the targeting of his Tehran office. The Iranian government subsequently announced a 40-day national period of mourning.
One day earlier, Trump had directed every federal agency to “immediately cease” all use of Anthropic’s technology. He characterized the company as “leftwing nut jobs” and claimed their stance was endangering “American lives.”
The Defense Department simultaneously classified Anthropic as a “supply chain risk” and unveiled a six-month timeline for eliminating its systems from military infrastructure. Defense Secretary Pete Hegseth stated that American service members would not be “held hostage by the ideological whims of Big Tech.”
Tensions between Anthropic and the Defense Department had been escalating over several months. Military officials demanded unrestricted access to Claude for any “lawful” defense application. Anthropic maintained its refusal, insisting it would not eliminate protections against domestic surveillance operations or attacks conducted without human authorization.
Anthropic declared on Friday that “no amount of intimidation or punishment from the Department of War will change our position.”
Claude had already become integrated into military infrastructure through collaborative agreements with Palantir and Amazon Web Services. The AI system was also utilized during the January mission that resulted in Venezuelan President Nicolás Maduro’s capture in Caracas.
OpenAI Moves In
Within hours of the Pentagon severing its relationship with Anthropic, OpenAI CEO Sam Altman revealed an agreement to integrate OpenAI’s AI models throughout the Defense Department’s classified network infrastructure.
Altman stated the partnership aligned with OpenAI’s core principles, which include bans on domestic mass surveillance and mandates for human oversight in weapons deployment. He additionally urged the Pentagon to extend identical terms to all artificial intelligence companies.
OpenAI representatives declined to confirm whether its technology would serve as a direct replacement for Anthropic’s defense department operations.
The Pentagon had previously executed multi-year artificial intelligence contracts valued at up to $200 million each with multiple firms, including Anthropic, OpenAI, Google, and Elon Musk’s xAI.
Funding and Rivalry
OpenAI revealed a $110 billion fundraising round on Friday, establishing the company’s valuation at $730 billion. Anthropic had completed a $30 billion capital raise in February.
Both organizations are advancing toward potential initial public offerings, with launches possible within the current year. Anthropic CEO Dario Amodei previously served at OpenAI before departing in 2020 due to disagreements about safety protocols being deprioritized.
Artificial intelligence specialists indicate that completely replacing Claude throughout military infrastructure will require several months, considering the technology’s deep integration with partners such as Palantir.


