As the U.S. military executes a blistering campaign of 1,000 strikes against Iranian targets, a secondary war is being waged in the boardrooms of Silicon Valley.
The integration of Anthropic’s Claude into the military’s Maven Smart System has turned weeks of traditional battle planning into real-time operations, but a widening rift between the tech giant and the Trump administration has opened the door for its chief rival, OpenAI.
For the past year, Anthropic has enjoyed first-mover advantage in the war theater. Paired with Palantir’s data-mining capabilities, Claude has matured into an essential tool for Central Command (CENTCOM). According to sources familiar with the system, the AI has been instrumental in identifying targets, issuing precise coordinates, and conducting post-strike evaluations. Its success in the current Iran operations follows its previous deployment in counter-terrorism plots and the high-profile capture of Venezuelan President Nicolás Maduro.
However, the relationship between the Pentagon and Anthropic has collapsed.
Hours before the first bombs fell on Iran, President Trump announced a ban on Anthropic’s tools, citing a bitter fight over the company’s refusal to allow its technology to be used for mass domestic surveillance and fully autonomous weaponry.
While Anthropic CEO Dario Amodei has publicly defended the use of AI to protect democracies, his moral red lines have created a strategic vacuum that the administration is eager to fill.
As the Pentagon begins a six-month phase-out of Anthropic, OpenAI is moving aggressively to seize the mantle. While Anthropic grapples with federal bans, OpenAI — alongside Elon Musk’s xAI — recently signed landmark agreements to work on classified government systems.
The timing is surgically precise.
OpenAI is currently rolling out GPT-5.3 Instant, a model specifically engineered to eliminate the “preachy” or “overly cautious” tone that has frustrated military and government users, it said in a blog post. By reducing unnecessary refusals and moralizing preambles, OpenAI is positioning its models as the pragmatic, compliant alternative to Anthropic’s ethically-focused Claude.
The stakes of this corporate competition are measured in life. The 18th Airborne Corps reported that AI integration allowed an artillery unit of just 20 people to perform the work of 2,000. Military commanders have become so dependent on these “videogame-like abilities” that the administration has signaled it will use government powers to forcibly retain Anthropic’s technology until a replacement, likely an OpenAI variant, is fully operational.
“We’re not going to let [Amodei’s] decision-making cost a single American life,” one official told the Washington Post, highlighting the cold reality of the new arms race. While the ethics of “machine-speed” warfare remain a subject of intense debate, the Pentagon’s trajectory is clear: The future of American defense will be powered by the AI firm most willing to trade philosophical caution for tactical utility.
As Claude faces its sunset in the Situation Room, OpenAI’s consistently helpful and less restrictive models are waiting in the wings to take command.
