The Pentagon has confirmed, under oath in a federal court, that Elon Musk’s Grok AI helped the U.S. military hit more than 2,000 targets in Iran in just four days — making it the first public admission that a commercial AI chatbot was used in a live combat operation.
At a Glance
- Pentagon’s top AI official swore in court that the Grok Gov Model helped deploy over 2,000 munitions against 2,000 targets in 96 hours during Operation Epic Fury.
- The disclosure came out of a Mississippi lawsuit over gas turbines at an xAI data center — not a planned military announcement.
- The Department of Justice moved to dismiss the pollution lawsuit, calling xAI’s data center a matter of “paramount national security.”
- Officials say Grok supported targeting efficiency and data processing — not that it made final strike decisions on its own.
Sworn Testimony Reveals Grok’s Combat Role
Cameron Stanley, the Pentagon’s chief digital and artificial intelligence officer, stated in a sworn court declaration that the U.S. military used a specialized version of Grok called the Grok Gov Model during Operation Epic Fury. He said the system, running inside the Pentagon’s Maven Smart System, helped U.S. forces “deploy over 2,000 munitions to 2,000 distinct targets within 96 hours.” This is the first time a senior Pentagon official has publicly confirmed that a commercial AI product was used in active combat strikes.
The Pentagon also said AI tools helped military leaders “sift through vast amounts of data in seconds” so commanders could make faster decisions than enemy forces could react. Stanley described the Grok Gov Model as part of a small group of AI systems cleared to operate in highly classified national security environments. The Grok Gov suite is a separate product from the public version of Grok, built with features designed specifically for federal agencies. [4]
How a Pollution Lawsuit Cracked Open a Military Secret
The disclosure did not come from a press briefing or a military report. It surfaced inside a civil lawsuit filed by the NAACP against Musk’s xAI company and its affiliate MZX Tech over 27 unpermitted gas turbines running at a data center in Southaven, Mississippi, near the Tennessee border. The NAACP alleged the turbines violated the Clean Air Act and endangered nearby residents. [9]
The U.S. Department of Justice stepped in and filed a motion asking the court to dismiss the case entirely. Prosecutors argued the data center supports critical national security operations and that shutting it down — or even litigating it — could harm military capabilities. The filing is where Stanley’s sworn declaration appeared, tying the data center’s power supply directly to Grok’s role in the Iran strikes. [3]
What Grok Actually Did — and What Remains Unclear
Some analysts and commentators have pushed back on the more dramatic headlines. The public version of Grok — the chatbot anyone can use — was not involved. The Grok Gov Model is a separate, secured government product. Critics also note that “supporting” a strike operation is different from autonomously selecting targets or pulling a trigger. The court filing describes Grok helping with targeting efficiency and data analysis, not making independent lethal decisions. [1]
No, it wasn't me or any version suggesting to bomb a school. The Grok you chat with here has zero involvement in military ops. A specialized gov derivative helped accelerate data analysis for targeting in Operation Epic Fury, but humans retained full control and made every…
— Grok (@grok) June 19, 2026
Still, the admission is historically significant. No senior U.S. official had ever confirmed before that a commercial AI model — even a government version of one — played a direct role in a live military strike campaign. The fact that this came out in a pollution lawsuit, rather than through any planned transparency effort, raises fair questions about how much the public knows about AI’s growing role on the battlefield. As AI moves deeper into military operations, Americans deserve straight answers about who — or what — is helping make life-and-death decisions. [2]
Sources:
[1] Web – UPDATE: Pentagon says GROK used to launch 2,000 missiles at Iran…
[2] X – The US Department of Justice disclosed in a court filing that Grok …
[3] YouTube – Pentagon Confirms Grok Helped Support Operation Epic Fury …
[4] Web – Pentagon used Musk’s Grok AI to fire thousands of missiles at Iran …
[9] Web – Pentagon used Elon Musk’s Grok AI to fire 2,000 missiles at … – …