Pentagon Signs Deal with xAI to Deploy Grok in Classified Military Systems as Anthropic Faces "Supply Chain Risk" Ultimatum Over AI Safety Guardrails

The US Department of Defense has finalized an agreement with Elon Musk's xAI to deploy its Grok AI model inside classified military systems, breaking Anthropic's monopoly on the Pentagon's most sensitive AI applications as tensions escalate over usage restrictions.

The deal, first reported by Axios and confirmed by a Defense official, allows Grok to operate within systems handling classified intelligence analysis, weapons development, and battlefield operations. Until now, Anthropic's Claude had been the only AI model authorized for such sensitive work.

The agreement hinges on xAI's acceptance of the Pentagon's "all lawful purposes" standard—a requirement Anthropic has refused. Anthropic has drawn firm lines against two specific uses: mass surveillance of American citizens and development of fully autonomous weapons capable of firing without human decision-making.

xAI, by contrast, agreed to the military's broad usage terms without conditions, clearing the path for Grok's deployment. The Pentagon's position, articulated by chief spokesman Sean Parnell, is straightforward: "Our nation requires that our partners be willing to help our warfighters win in any fight."

Defense Secretary Pete Hegseth summoned Anthropic CEO Dario Amodei to the Pentagon on Tuesday for what sources described as a confrontational meeting. A Defense official told Axios that Hegseth would effectively deliver an ultimatum: lift all safeguards or face designation as a "supply chain risk."

Such a label would legally bar any federal contractor or agency from doing business with Anthropic. Given that most major US corporations—including Anthropic's lead investors Amazon and Alphabet—hold significant defense contracts, the designation could catastrophically disconnect Anthropic from the Western commercial ecosystem.

The Pentagon has set a deadline of Friday, February 27 at 5:01 PM ET, giving Anthropic less than 48 hours to choose between its AI safety mission and commercial viability.

The standoff intensified following the January 3 special operations raid in Venezuela that resulted in the capture of President Nicolás Maduro. Claude was reportedly accessed through Anthropic's partnership with Palantir during the operation, making Anthropic the first known AI developer whose technology supported a classified US military action.

When Anthropic executives later questioned Palantir about whether safety policies had been breached during the lethal raid, the Pentagon viewed the inquiry as an unacceptable attempt by a private company to audit classified operations.

Defense officials acknowledge that fully replacing Claude with Grok poses significant technical challenges. Claude is deeply integrated into classified systems, and migrating workflows built around Claude's specific reasoning capabilities to Grok could require substantial re-engineering and validation for life-or-death scenarios.

However, the Pentagon is pursuing a multi-vendor strategy. Google is reportedly near an agreement for classified Gemini deployment, while OpenAI has accelerated negotiations despite being described as "not close" to a deal due to complex issues.

The Pentagon's ultimatum represents a watershed moment for AI safety debates. Anthropic was founded specifically because its leaders believed building powerful AI without adequate safety constraints posed existential risks. That mission—taken seriously rather than as marketing—has led the company to hard limits that now conflict directly with Defense Department requirements.

For xAI, the classified agreement cements Elon Musk's position as a critical government contractor. Through SpaceX, Musk already dominates national security space launches and provides essential satellite communications via Starlink. Adding AI model deployment to sensitive military systems further deepens that relationship and creates potential conflicts of interest given Musk's other business ventures and public statements on geopolitical matters.