The global military AI landscape is undergoing a major upheaval. On February 24, 2026, according to cnBeta, Axios reported that Elon Musk's artificial intelligence company xAI has officially signed an agreement with the U.S. Department of Defense, allowing its large model Grok to be deployed in the most sensitive classified systems of the U.S. military.

Key Turnaround: Grok Fills the Gap Left by Anthropic

Previously, Anthropic's Claude was the sole AI pillar in the U.S. military's classified systems, used for handling weapons development and battlefield intelligence. However, recent serious disagreements have emerged between the two parties:

  • Disagreement on Principles: Anthropic firmly prohibited Claude from being used for mass surveillance or developing fully autonomous weapon systems.

  • Musk's "Straightforward" Approach: In contrast, xAI has agreed to the Pentagon's "usable for all legal purposes" standard, opening the door for military use in sensitive areas.

Breakthrough Moment: The Pentagon Threatens Sanctions Against Anthropic

The Pentagon has taken a strong stance against Anthropic's "non-cooperation":

  • "Last Chance": Defense Secretary Pete Hegseth plans to meet with Anthropic CEO Dario Amodei this week. If the company does not lift its usage restrictions, the Pentagon may label it as a **"supply chain risk"** and impose penalties.

  • Replacement Cost: Although the Pentagon has started looking for alternatives, officials have admitted that unloading and replacing the deeply embedded Claude in critical missions is an extremely complex task.

Corporate Rivalry: Gemini and ChatGPT Accelerate Their Entry

The ticket to enter the classified system is sparking a collective hunt by Silicon Valley giants:

  • Google Gemini: It is reported to be "close" to reaching an agreement for deployment in the classified system.

  • OpenAI: Although already used by the U.S. military in non-classified environments, its progress in entering the classified system remains slow.

  • Unified Standards: Officials from the Biden administration have insisted that all models entering the classified system must adhere to the "full-use" standard, which undoubtedly places AI vendors in the ultimate dilemma between "technical ethics" and "government contracts."

Industry Insight:

Musk's xAI has seized the opportunity with a more flexible (or more radical) compliance attitude, not only significantly boosting its position in the military AI field but also signaling that future AGI competitions will inevitably be deeply tied to national security and weaponization.