Artificial intelligence has long sat at the center of heated debates about future warfare, but a newly revealed US government filing suggests the future may already be here.
According to legal documents reviewed by Agence France-Presse (AFP), the US government has acknowledged that a version of Elon Musk’s Grok AI technology was used within military operations tied to strikes against Iran. The disclosure surfaced in a court fight over power generation equipment at xAI data centers.
The filing, submitted by the US Department of Justice on June 15, was intended to defend gas turbines that power an xAI facility currently facing an environmental lawsuit.
The US government argued that the case has implications beyond environmental concerns, stating that it “threatens American national, economic and energy security by seeking to shut off the power supply for artificial intelligence innovation that supports the Department of War’s military operations.”
Project Maven enters the spotlight
To support its case, the government submitted sworn testimony from Pentagon AI chief Cameron Stanley, who said a specialized system called the “Grok Gov Model” is already operating within Project Maven.
Project Maven is the US military's AI-assisted targeting and intelligence initiative. It was previously powered by Anthropic's Claude AI model before changes in government partnerships shifted the project toward other providers.
In Stanley’s testimony, he said Maven Smart Systems “enabled US forces to deploy over 2,000 munitions to 2,000 distinct targets within 96 hours during Operation Epic Fury.” Stanley also highlighted “the greatly increased operational efficiency made possible by the Grok Gov Model.”
The filing did not describe Grok as independently making battlefield decisions. Instead, the documents connect the AI system to broader targeting and operational support within the military program.
Environmental lawsuit unexpectedly exposes military connection
The details became public through a lawsuit brought by the NAACP against xAI.
The civil rights organization alleges that xAI has been operating dozens of gas turbines without permits, violating provisions of the Clean Air Act and affecting predominantly Black communities near the facility.
xAI disputes those claims and argues that the turbines are temporary and mobile systems that do not fall under the same regulatory requirements. The Justice Department argued that restricting power to the facilities could affect systems supporting military activity, bringing national security concerns directly into a case initially focused on pollution claims.
AI companies face growing pressure over military work
The government filing also sheds light on how xAI secured a role in one of the Pentagon's most sensitive AI projects.
According to the documents, the US government ended its contracts with Anthropic at the end of Feb. after the company refused to allow its AI systems to be used for fully automated strikes or the mass surveillance of Americans. Following that decision, the Pentagon turned to other AI companies, including Google, OpenAI, and xAI. The move has not been without controversy.
More than 600 Google employees have previously called on the company not to provide AI technology for classified military operations, while critics have warned that the growing use of AI in warfare raises serious ethical and security concerns.
Despite the shift, the transition has been gradual. The government acknowledged in March that Anthropic's Claude model was still being used in operations linked to the conflict with Iran. The latest disclosure, however, suggests Musk's Grok has now become a key component of the Pentagon's expanding AI arsenal.
Also read: Elon Musk’s reported trillionaire milestone after SpaceX’s IPO raises new questions about how his space, AI, and defense interests are converging.