Partly. Outdated intelligence compounded AI overreliance

The strike and the faulty data are well documented. The data was the government's, and the targeting failure is mainly described as human and process failure: stale intelligence and reduced civilian-harm review. Reporting on the Pentagon review says Maven sped up target-list work and that staff over-relied on it, so the Palantir tool plausibly contributed. Palantir disputes this, and the final Pentagon report has not been released. The claim also names Anthropic. Claude is reported as part of the wider Maven ecosystem, but its role in this specific strike is not established. The claim is therefore only partly supported: it is accurate on faulty data and on Palantir's tool being involved, but it overstates how far technology rather than human decisions was the cause, and it is unproven for Anthropic.

• Two Tomahawk missiles hit Shajarah Tayyebeh Elementary School in Minab on 28 February 2026, killing at least 123 children.
• The site was cataloged as an IRGC facility from outdated data, and Maven returned it as a recommended day-one target (Bloomberg, 18 September 2026, citing the unreleased Pentagon review).
• Investigators reportedly found that some staff expected Maven to flag stale records, which it was not designed to do. CENTCOM's civilian-harm team had been cut from ten people to one.
• Palantir says it is not responsible for the underlying data and that there is no evidence its software was at fault. Its CTO said in October 2026 that human error was the likely cause.
• Reports say Claude works with the Maven Smart System, but no source I found shows Claude selected or validated this target. The Pentagon investigation is not final.

1. Pentagon probe cites Palantir Maven in Iran school strike aiweekly.co
2. Pentagon Investigators Say Overreliance on Palantir AI Tech Contributed to U.S. Strike That Killed 123 Iranian Children gizmodo.com
3. Inside US Military ‘Kill Chain’ That Destroyed an Iranian School bloomberg.com
4. Pentagon Rewires Palantir Maven After Iran School Strike aiweekly.co
5. Palantir CTO Says ‘Human Error’ Likely Caused Deadly Iran Strike That Killed 123 Children — Disputes Clai - Benzinga benzinga.com
6. Palantir’s answer to a deadly targeting failure was more AI - TheStreet thestreet.com
7. 2026 Minab school attack en.wikipedia.org
8. Palantir Hits Records as Maven Scrutiny Grows finance.yahoo.com

The claim has a supported core: faulty military intelligence and reliance on Palantir’s software reportedly contributed to the strike. But it compresses several distinct failures into a single explanation and assigns Anthropic a causal role that the cited reporting does not establish. Humans still reviewed and authorized the target; investigators reportedly identified failures in both intelligence processing and human verification, rather than an autonomous AI decision to bomb a school.

• Shajareh Tayyebeh elementary school in Minab was struck on February 28, 2026; preliminary U.S. findings attributed the strike to American forces using outdated intelligence.
• The Defense Intelligence Agency’s information reportedly still classified the school site as an Iranian Revolutionary Guard naval facility.
• In September 2026, Bloomberg’s investigative reporter said investigators found that personnel overrelied on Maven, which failed to flag outdated imagery and inconsistent intelligence.
• Palantir’s Maven incorporates information from language models such as Anthropic’s Claude, but this general integration does not demonstrate Claude’s involvement in this particular targeting error.
• On October 11, 2026, Palantir’s CTO disputed Bloomberg’s account, attributing the likely cause to human error; the final report remained unreleased.

1. theguardian.com
2. transcripts.cnn.com
3. semafor.com
4. bloomberg.com
5. factcheck.org
6. washingtonpost.com
7. macmyths.com
8. en.wikipedia.org

On February 28, 2026, during the opening strikes of Operation Epic Fury, U.S. Tomahawk missiles destroyed the Shajareh Tayyebeh elementary school in Minab, Iran. The location had historically been part of an adjacent Islamic Revolutionary Guard Corps base, but satellite imagery and local records showed it had been partitioned off as a civilian school years earlier. In September 2026, media investigations detailing an internal Pentagon review revealed that U.S. Central Command planners used Palantir's Maven Smart System (which integrated Anthropic's Claude for tasks such as ranking targets) to process hundreds of strikes in compressed timeframes. Planners fed outdated database entries into the system and erroneously assumed the AI would cross-check or flag stale records. While the AI tools accelerated the strike package, officials and Palantir emphasized that the disaster stemmed from a failure of human verification, staffing cuts to civilian protection teams, and flawed base data, rather than the AI software autonomously deciding to bomb the site.

• On February 28, 2026, a U.S. missile strike hit the Shajareh Tayyebeh elementary school in Minab, Iran, killing over 150 people, mostly children.
• The building had once been part of an adjoining military compound years earlier, but had been walled off and operated openly as a school for several years.
• Target planners utilized the Maven Smart System, built by Palantir and integrating Anthropic's Claude model, to accelerate day-one strike packaging.
• A Pentagon investigation reported in September 2026 found military personnel overrelied on Maven to catch errors, expecting it to flag stale intelligence records that it was not designed to detect.
• Palantir stated that its software was not at fault and was not responsible for the military's underlying intelligence data deficiencies.

1. militarytimes.com
2. gizmodo.com
3. remio.ai
4. yahoo.com
5. theguardian.com
6. tradingview.com

The question asks whether a specific Iranian school bombing was caused by Anthropic and Palantir technology using faulty data. While both companies have defense-related AI partnerships, no reliable evidence ties any such system to a school strike in Iran or to a documented faulty-data error causing it. Without verified reporting or investigation, the causal claim is not shown.

• Anthropic and Palantir announced a partnership in 2024 to make Claude AI available on Palantir's platforms for defense and intelligence customers.
• Palantir provides data analytics and AI tools to governments and militaries, but it does not publicly disclose specific targeting operations.
• No independent investigation has confirmed that an Iranian school was struck as a result of an AI system's faulty data.
• Attributing a specific airstrike to an AI targeting error requires evidence that the system was used and that its output directly caused the strike.
• The claim appears to conflate the existence of defense AI partnerships with a specific, unverified battlefield incident.

The claim that the Iranian school was bombed because of tech built by Anthropic and Palantir using faulty data is not supported by available evidence. Multiple investigations and reports, including from Bloomberg and Wikipedia, confirm that the school was misclassified as a military target due to a combination of overreliance on AI tools, human error, and the rapid pace of the military operation. Palantir explicitly stated that it is not responsible for the underlying data or intelligence deficiencies, and there is no evidence that its software or Anthropic’s Claude model used faulty data in this incident. The related harm was misclassification, but not due to faulty data from these companies' systems.

• The U.S. strike on the Shajareh Tayyebeh school in Minab, Iran, on February 28, 2026, killed 156 people, including 120 children.
• Palantir’s Maven Smart System, which integrates AI for military targeting, was used in the strike, and Anthropic’s Claude model was part of the U.S. military’s AI infrastructure at the time.
• Palantir stated that the company is not responsible for the underlying data or identifying intelligence deficiencies, and there is no evidence its software was at fault in the Minab strike.
• Investigations found that the school was misclassified as a military target due to overreliance on AI tools, human error, and the rapid pace of the opening assault, not because of faulty data from Anthropic or Palantir.
• The U.S. Department of Defense had ceased using Anthropic products, including their removal from Palantir’s Maven Smart System, by September 2026.

1. Anthropic en.wikipedia.org
2. Inside US Military ‘Kill Chain’ That Destroyed an Iranian School bloomberg.com
3. Anthropic–United States Department of Defense dispute en.wikipedia.org
4. Overreliance On Palantir AI Technology Played Role In Iran School Bombing: Report - Yahoo News Malaysia malaysia.news.yahoo.com
5. Claude (AI) en.wikipedia.org
6. The Algorithm of War: How Palantir and Anthropic Became Co-Belligerents in America’s Iran Campaign | by HILAL AHMAD KHAN | Medium medium.com
7. 2026 Minab school attack - Wikipedia en.wikipedia.org
8. Reporting Confirms: Stupid Over-Reliance On Palantir AI Helped Lead To US Bombing Of Iranian Schoolgirls | Techdirt techdirt.com