Donald Trump took to his Truth Social platform to praise Palantir Technologies on Friday. The software company, he wrote, “has proven to have great war-fighting capabilities and equipment. Just ask our enemies.”
It was a big endorsement for the company, considering it came directly from the president, and behind it is the story about how artificial intelligence has moved from writing your emails to the centre of warfare.
The Iran war is probably the world’s first large-scale example of artificial intelligence directing military operations in real time. And at the centre of it all is a software company that most people outside Washington and Wall Street hadn’t heard of until recent years.
Before bombs fell on Tehran on February 28, at the Pentagon’s direction, a software sifted through mountains of satellite imagery and drone footage to prepare more than a thousand strike options for American commanders.
Decisions that might have taken human analysts days were compressed into minutes. The strikes themselves were still carried out by people, but Maven did the heavy lifting in between.
At the heart of this was Palantir’s Maven Smart System (MSS), which runs on Claude, the artificial intelligence platform developed by Anthropic.
While Claude had previously been used in counter-terrorism work and in the kidnapping of Venezuelan president Nicolas Maduro, this is “the first time it has been used in major war operations,” the Washington Post reported.
According to the Post, over the past year, the government allowed the Maven and Claude combination to develop into a tool that is now in daily use across most branches of the US military.
Rear Admiral Brad Cooper, who leads American military operations in the Middle East, has spoken publicly about how AI platforms are helping officers sift through enormous volumes of information in seconds, giving leaders the ability to “cut through the noise and make smarter decisions faster than the enemy can react.”
What Is Maven, And How Does It Work?
Maven Smart System is a command-and-control technology. Its origins trace back to a 2017 Pentagon programme called Scarlet Dragon, conceived by the US Army’s 18th Airborne Corps. The goal was to create a unified, AI-powered network, known as a Joint All-Domain Command and Control System, that brings together intelligence streams from every branch of the military: Army, Navy, Air Force, and even Space.
In practical terms, Maven identifies, tracks, and classifies objects on the battlefield, whether vehicles, weapons, or buildings, and dramatically reduces the time it takes to identify a target from several hours to under a minute. It draws on data from drone footage, satellite sensors, and other sources. Beyond target identification, Maven is also designed to assist with what military planners call predictive logistics, helping commanders forecast where supplies will be needed before shortages occur, and supporting decisions about how and when to deploy firepower.
Admiral Brad Cooper, head of US military operations in the Middle East, has praised AI systems for allowing officers to “sift through vast amounts of data in seconds, so our leaders can cut through the noise and make smarter decisions faster than the enemy can react.”
How Claude Makes Maven Think
Anthropic and Palantir joined forces in November 2025 and announced that they would integrate Claude, Anthropic’s large language model, into the Maven platform.
Claude’s role within Maven is to make sense of the data the system collects. Rather than simply flagging objects or locations, the integrated system can contextualise information, weigh it against historical patterns, and present commanders with assessments that go beyond what raw imagery or sensor data would offer on their own.
The arrangement has not been without friction. In February, however, the Pentagon announced that it intended to phase out Claude from the Maven system after Anthropic declined to permit its AI to be used in autonomous weapons or surveillance applications. Palantir said it could find alternatives.
The Pentagon Makes It Official
Last month, Reuters reported that the Pentagon had formally designated Maven as an “official programme of record,” a classification that signals long-term integration across the US military. In a letter obtained by the news agency, Deputy Defence Secretary Steve Feinberg said the platform would give commanders “the latest tools necessary to detect, deter, and dominate our adversaries in all domains.”
But Maven is not new to combat. The system, in various earlier forms, has been used by the American military for roughly nine years. It was first deployed in operations against the Islamist militant group al-Shabaab in Somalia. It was used to plan strikes in Iraq and Syria in February 2024. It supported a civilian rescue operation in Afghanistan in 2021 and has been involved in monitoring efforts connected to Russia’s ongoing war in Ukraine. In April last year, NATO signed a deal with Palantir to purchase Maven, suggesting the technology is not going away.
The Criticism
The use of AI in the Iran campaign has not gone unquestioned. Analysts have warned that tools like Maven leave very little time for meaningful verification of the output they produce, raising the risk of incorrect targets being struck. If the AI gets it wrong, the human in the loop may not have the opportunity, or the information, to catch the error before it becomes irreparable.
In the US Congress, senior Democrats have called for greater scrutiny of AI platforms like Maven. The anti-AI rhetoric has sharpened following reports about a US strike on a school in the Iranian town of Minab, in which Iranian officials say 168 people were killed, including approximately 110 children, on the opening day of the war. Pentagon officials have faced direct questions about whether AI tools were used to identify that target.
Louis Mosley, Palantir’s head for the UK and Europe, acknowledged that platforms like Maven had been “instrumental” to the US management of the Iran war, but insisted that responsibility for how their output is used must remain with the military.
“There’s always a human in the loop. So there is always a human who makes the ultimate decision. That’s the current set-up,” he said in an interview with the BBC this month.
Palantir’s software has become so deeply embedded in the Pentagon that the company has graduated from being a niche defence contractor to a central figure in the US military operations.
In all this, there’s a substantial victory for Palantir. The company has secured a growing run of government contracts, including a deal with the US Army announced last year, worth up to $10 billion. The company’s share price has roughly doubled over the past year. Its market capitalisation now sits at approximately Rs 3 trillion (350 to 360 billion US dollars).