The Pentagon turned to Grok to fire 2,000 missiles at Iran

A high-ranking official for the US Department of War has confirmed that the United States used xAI’s chatbot Grok to fire 2,000 missiles during the war with Iran.

Cameron Stanley, Chief Digital and Artificial Intelligence Officer (CDAO) at the Pentagon, submitted a sworn statement to the US District Court Northern District of Mississippi to defend xAI’s use of gas turbines for its Colossus 2 data center.

The Colossus 2 data center is used to train and upgrade future Grok models, including the one that’s being used by the Pentagon.

NAACP, a civil rights organization, has filed a lawsuit against xAI to force the company to stop using the gas turbines powering the Colossus 2 data center. The organization claims that xAI built the turbines without the proper permits and that they are detrimental to the health of nearby residents.

“This action directly threatens our ongoing national security interests,” Stanley says in his statement. He claims that Grok is one of 4 frontier AI models that are capable of supporting national security applications.

The US government uses a specialized Grok model, called Grok Gov, to execute military operations. It’s capable of gathering intelligence, selecting military targets, and conducting predictive analysis for logistics, personnel management, and medical supply lines.

During the war with Iran, Grok Gov was used to deploy 2,000 missiles to 2,000 distinct targets within 96 hours during Operation Epic Fury, which Stanley calls “a testament to the greatly increased operational efficiency.”

Shutting down the gas turbines that power xAI’s available infrastructure, as is requested by the plaintiffs, would be catastrophic. According to Stanley, this would mean that xAI loses its capacity to train and develop future versions of Grok. In addition, it would impede the Pentagon’s ability to secure the United States’ national security interests.

“Therefore, preserving xAI’s current data center capacity is a matter of paramount national security,” Stanley explains.

Furthermore, the Pentagon relies on xAI as a primary classified compute infrastructure provider. Any impediment would diminish xAI’s capacity to support current military operations.

“If Grok Gov Model cannot be deployed, refined, and upgraded across the DoW [Department of War, ed.] due to either limitations in energy supply or limited reserve compute capability, such as those requested by plaintiffs in this matter, the many tools deployed by military and civilian personnel alike, which rely on Grok Gov Models, would be severely impacted,” the high-ranking official concludes his statement.

While the US government relies on several AI companies to conduct military operations, the Pentagon is also embroiled in a legal battle over the use of another AI tool used for warfare.

Back in February, US President Donald Trump ordered government officials to immediately stop using Anthropic’s AI technology, as it was designated as a supply-chain risk.

“Anthropic delivered a master class in arrogance and betrayal as well as a textbook case of how not to do business with the United States Government or the Pentagon. The Department of War must have full, unrestricted access to Anthropic’s models for every LAWFUL purpose in defense of the Republic,” Secretary of War Pete Hegseth said at the time on X.

More recently, Anthropic disabled Fable 5 and Mythos 5 after concerns that they could be used by foreign military intelligence services in China, Russia, and other countries.