Skip to content
News Incident

Pentagon official states Grok Gov directed 2,000 munitions in 96 hours during Operation Epic Fury

· by Pondero Newsdesk

The short version

The DOD's chief digital and AI officer swore in federal court that Grok Gov enabled U.S. forces to deploy over 2,000 munitions to 2,000 distinct targets in 96 hours. The statement is the first on-record U.S. government admission that a commercial LLM supported live kinetic targeting at scale.

Pentagon official states Grok Gov directed 2,000 munitions in 96 hours during Operation Epic Fury

Cameron Stanley, the Department of Defense's chief digital and artificial intelligence officer, stated in a June 16, 2026 sworn court filing that xAI's Grok Gov "enabled U.S. forces to deploy over 2,000 munitions to 2,000 distinct targets within 96 hours during Operation Epic Fury." The filing is the first confirmed on-record admission by a U.S. government official that a commercial large language model supported live kinetic targeting at operational scale.

What happened

Stanley's statement appeared as a declaration in NAACP v. xAI, the federal environmental lawsuit in the Northern District of Mississippi challenging xAI's operation of at least 57 unpermitted gas-burning turbines at its Colossus 2 data center in Memphis. The Trump administration filed the declaration to support a motion to dismiss the case on national security grounds.

In the filing, per the court record on CourtListener, Stanley described Grok Gov as "a matter of paramount national security" and listed it as one of four AI models "currently capable of supporting national security applications." The specific targeting claim, 2,000 munitions to 2,000 distinct targets in 96 hours, appeared in the same declaration and was cited to Operation Epic Fury, the U.S. military operation conducted as part of the 2026 U.S.-Israel-Iran conflict.

The New Republic reported the story June 17, noting it was the first time the Trump administration admitted to using xAI's Grok in the Iran war. The DOD awarded xAI a $200 million federal contract to install Grok for Government into its systems, per New Republic's earlier reporting. That contract predated the Operation Epic Fury deployment.

Why it matters

The court filing did not describe how Grok Gov was integrated into the targeting workflow, what human oversight layers existed, or what role the model played relative to conventional intelligence and weapons guidance systems. Stanley's declaration asserted the outcome but provided no technical methodology, per the public court record.

The admission is significant for several reasons. It names a commercial AI vendor product by name in a targeting context at a scale not previously confirmed on the record by any U.S. official. It came as a litigation argument rather than a deliberate policy disclosure, raising questions about what other AI targeting use cases may be in operational use without public disclosure. The DOD has not issued a separate statement elaborating on the claim or clarifying the role Grok played versus human decision-makers in the loop.

The NAACP's lawsuit, which concerns Clean Air Act violations from xAI's turbines, now sits at an unusual intersection of environmental regulation and claimed national security necessity. A court ruling that accepts Stanley's national security framing could effectively insulate commercial AI infrastructure from environmental enforcement when the government asserts a military dependency.

Context

Grok's reliability record adds friction to the assertion. The New Republic's reporting documented that researchers had found Grok spreading false content during the Iran war, including misidentifying strike footage and giving contradictory answers to verification queries. Other government agencies had flagged Grok as a security risk in earlier internal assessments.

The Pondero Newsdesk separately covered the early stage of this case in June 2026, when the DOJ signaled it might intervene on xAI's behalf on AI policy grounds. Stanley's sworn statement represents the Trump administration's escalation from potential intervenor to active declarant, and the first time the Grok-targeting connection has been stated by a named official under oath.

Stanley listed the other three AI models he considered capable of supporting national security applications only in general terms in the public filing. No other vendors were named.

What to watch next

Congress has not yet announced hearings in response to the disclosure. The next formal step in NAACP v. xAI is the court's ruling on the national security dismissal motion. If the court declines to dismiss, xAI faces the NAACP's preliminary injunction request on the turbine operations. Whether the DOD issues any clarifying statement about Grok Gov's targeting role, the oversight chain, or compliance with the department's 2022 AI ethics principles will determine whether this filing becomes a policy moment or remains a litigation footnote.

Sources