Skip to content
Submit incident
AII-1000-92217
ConfirmedA court, a regulator or the organization itself has stated the core facts, or the record cites a court or regulator source.

Pentagon reveals Grok AI used in Iran missile strikes in court filing

The U.S. Pentagon revealed in a June 2026 court filing that xAI's Grok AI system, specifically a derivative called Grok Gov Model, was deployed in Project Maven to assist in targeting decisions that resulted in over 2,000 munitions being deployed on 2,000 distinct targets in 96 hours during military operations against Iran. The government disclosed this information while defending xAI against a pollution lawsuit, arguing that disrupting the company's data center infrastructure would threaten national security.

15 June 2026 · xAI · Security and cyberattacks
15 June 2026
AII-1000-922172 sources ↓
Quick facts
Category
Security and cyberattacks
Country
United States
Source
BFMTV
Date
15 June 2026
Drafted by
Raidu Incident Detection

What happened

The U.S. Department of Justice filed a court memorandum on June 15, 2026, revealing that Grok, an AI model developed by xAI, was used in military strikes against Iran.

Cameron Stanley, the Pentagon's AI chief, testified under oath that a derivative tool called 'Grok Gov Model' is deployed within Project Maven, the military's AI-assisted targeting program. Stanley stated that Maven processes enabled U.S. forces to deploy over 2,000 munitions on 2,000 distinct targets in 96 hours during the war against Iran.

The government disclosed this information in response to a lawsuit filed by the NAACP against xAI over unlicensed turbines operating at a data center near Memphis that power the Colossus 2 supercomputer used to train Grok.

Reported impact

Stanley testified that Maven users consume approximately 2 billion tokens per day, equivalent to up to 6 million pages processed, making xAI's computing infrastructure indispensable according to the Pentagon official.

Index analysis

The Pentagon argued that the pollution lawsuit 'threatens the national, economic and energy security' of the United States by risking disruption to AI infrastructure now used by the military.

Project Maven previously relied on Anthropic's Claude model but the government ended contracts with Anthropic in late February after the company refused to allow use of its outputs for certain military purposes.

Classification

Category
Security and cyberattacks
Industry
Government, security and defence
Country
United States
Harm status
Actual harm
Intent
Unknown
Timing
Unknown
Harm types
Physical death
AI systems
Grok, Grok Gov Model

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

BFMTV
"Plus de 2.000 munitions sur 2.000 cibles distinctes": pour défendre l'IA Grok d'Elon Musk visée par une plainte, le Pentagone révèle qu'elle a servi dans les frappes en Iran
17 June 2026
lindependant.fr
Guerre au Moyen-Orient : Grok, l'IA d'Elon Musk, a servi dans les frappes en Iran, révèle le Pentagone
17 June 2026