Elon Musk's artificial intelligence company, Grok, was used in the attacks against Iran, revealed the US Administration in a judicial document.
© Lusa
17/06/2026
by
Lusa
A disclosure is contained in a document of the Department of Justice of the United States, which defends the gas turbines of a gigantic data center of the multimillionaire company, the xAI, the target of an environmental complaint.
In the document, submitted on 15 June and consulted by the agency France-Presse (AFP), the Department of Justice argues that the complaint “threats national, economic and energy security” of the country by risking cutting energy supply to AI infrastructure now used by the army.
To support the argument, the department presents the testimony of Cameron Stanley, responsible for the IA at the Pentagon, who declares under oath that a tool derived from Grok, the “Grok Gov Model”, is already being used within the framework of the “Project Maven”, the Army’s AI-assisted target selection program, initially based on the Claude model of Anthropic.
According to the statement, Maven’s lawsuits “allowed US forces to launch more than 2,000 ammunition against 2,000 separate targets in 96 hours” during the war against Iran. The senior employee sees in these numbers “the proof of a very significant increase in operational efficiency, made possible by the Grok Gov Model”, without specifying whether this is the only model used by the program.
Stanley also states that Maven users consume “about two billion tokens” (computing units) “per day”, equivalent to “up to six million pages” processed, a volume that, according to the person responsible, makes the xAI computing infrastructure indispensable.
The turbines targeted by the complaint feed the Colossus 2, a supercomputer of the xAI that feeds the Grok, outside Memphis.
The NAACP (National Association for the Progress of Persons of Color) sued the xAI, accusing it of exploring dozens of turbines without license, in violation of the law on air quality.
Turbines pollute mostly black neighborhoods, says the association. The xAI argues that its turbines are temporary and mobile and are therefore not subject to the regulation mentioned.
In late February, Donald Trump's Administration terminated contracts with Anthropic, which refused to use its tools for fully automated attacks or mass surveillance by Americans.
The Pentagon then resorted to Google, OpenAI and xAI, the other three leading American AI companies. The transition, however, took time and Washington had to admit in March that Claude continued to be used in the war in Iran.
The military use of the AI raises controversy. At Google, more than 600 employees demanded in April that no AI be provided to the army for confidential operations. The group had already abandoned the Maven project in 2018, under pressure from its engineers.
Close to Donald Trump, whom he supported and advised, Musk joined the XAI in his SpaceX space group in February, which held, on June 12, the largest entry in the stock exchange in history.

