Pentagon investigators have made a shocking revelation about a 28 February strike that killed 123 children at an Iranian school, blaming it on flawed intelligence, outdated imagery and over-reliance on Palantir's Maven Smart System.
Two Tomahawk cruise missiles struck the Shajarah Tayyebeh Elementary School in Minab, Iran, on 28 February. More than 150 people died, including at least 123 children, according to an investigation by Bloomberg. Pentagon investigators found that flawed intelligence, outdated imagery and over-reliance on AI were among the contributing factors.
The internal review, based on interviews with officials involved in the investigation, found that the strike was not the product of a single catastrophic error. Instead, investigators identified a chain reaction as an AI-assisted targeting system accelerated parts of a process that had already been compressed by the pace of the campaign.
How Compressed War Planning Bypassed the Kill Chain
US officials involved in the investigation said the campaign called for more than 1,000 targets to be struck within the first 24 hours, dramatically compressing the time available to assess individual targets. Bloomberg reported that work that had previously taken hours was condensed into minutes in parts of the targeting process.
The school itself occupied ground that complicated the targeting process. The site had once been part of an Islamic Revolutionary Guard Corps facility, but the school had been built on the site and separated from the military compound by walls and entrances.Commercial satellite imagery showed changes to the site years before the strike, including features associated with a functioning school.
Into that compressed process came Maven Smart System, a commercial AI platform developed by Palantir and used by the US military for intelligence analysis and operational decision support. Maven can fuse large numbers of data inputs to help users coordinate military operations, including targeting and command and control.
According to officials involved in the Pentagon investigation, some US Central Command personnel expected Maven to flag stale information or inconsistencies in the intelligence assembled for potential targets. Investigators found that expectation was misplaced.
The Pentagon had catalogued the site as an IRGC facility, and that classification was not corrected before the strike.
Palantir disputed any characterisation that its software bore operational responsibility. A company spokesperson said Palantir 'is not responsible for the underlying data nor identifying intelligence deficiencies' and said there was no evidence its software was at fault in the Minab strike.
Bloomberg also reported that, after the strike, Palantir built new Maven capabilities to re-review underlying intelligence for factors that could disqualify a target and to flag inconsistencies and inaccuracies that human review may have missed.
What Minab Investigation Reveals About AI in Kill Chain
The military's kill chain is the sequence of steps used to move from intelligence and target identification to weapons release. Traditionally, the process involves intelligence analysts, imagery specialists, targeteers, lawyers, operational commanders and launch crews. Bloomberg's reconstruction found that Maven had become embedded between initial intelligence inputs and later stages of review, while parts of the process were compressed from hours into minutes.
Within hours of the 28 February attack, some Pentagon personnel knew the United States was responsible, according to officials involved in the internal investigation. The Pentagon has not publicly released the full findings of that investigation, and officials have continued to face questions about how the target was selected and reviewed.
