Pentagon reviews Anthropic AI contract after Maduro raid raises supply chain risk concerns
Subscribe to our free newsletter today to keep up to date with the latest supply chain industry news.
A dispute over the possible use of Anthropic’s artificial intelligence model during the US operation targeting Venezuelan leader Nicolás Maduro has prompted a Pentagon review of the company’s contract, with senior officials raising concerns that the firm could represent a supply chain risk.
Axios first reported on tensions between the Pentagon and Anthropic, a technology company known for emphasizing safeguards on advanced AI systems.
Anthropic secured a 200 million dollar contract with the Pentagon in July 2025. Its AI model, Claude, became the first such system deployed within classified Defense Department networks.
Now, “the Department of War’s relationship with Anthropic is being reviewed,” chief Pentagon spokesman Sean Parnell told Fox News Digital.
“Our nation requires that our partners be willing to help our warfighters in any fight,” Parnell said.
Questions over AI use in Maduro operation
According to a senior administration official, tensions escalated when Anthropic inquired whether Claude had been used during the raid to capture Maduro. The question raised concerns within the department that the company might object to certain operational uses of its technology.
“Given Anthropic’s behavior, many senior officials in the DoW are starting to view them as a supply chain risk,” a senior War Department official told Fox News Digital. “We may require that all our vendors and contractors certify that they do not use any Anthropic models.”
The official said an Anthropic executive posed the question to an executive at Palantir, the company’s partner in Pentagon contracting. The Palantir executive alerted the Pentagon, saying he was concerned that the inquiry suggested Anthropic might disapprove of its software being used in that operation.
Palantir did not immediately respond to a request for comment.
Anthropic responds to Pentagon concerns
Anthropic disputed that account. A spokesperson said the company “has not discussed the use of Claude for specific operations with the Department of War” and has not addressed such matters with industry partners beyond routine technical discussions.
The spokesperson added that Anthropic’s conversations with the Pentagon have focused on usage policy questions, specifically the company’s limits regarding fully autonomous weapons and mass domestic surveillance. Those issues, the spokesperson said, do not relate to current operations.
“We are having productive conversations, in good faith, with DoW on how to continue that work and get these complex issues right,” the spokesperson said.
Pentagon officials denied that restrictions tied to mass surveillance or fully autonomous weapons are central to the current dispute. Instead, the department has pressed leading AI firms to authorize their models for all lawful purposes, seeking assurances that commercial systems can be deployed in sensitive operational environments without company imposed limitations.
Defense department push for unrestricted lawful AI use
A senior War Department official said other major AI companies are cooperating with the Pentagon to ensure their models are available for lawful military use.
“OpenAI’s ChatGPT, Google’s Gemini and xAI’s Grok have all agreed to this in the military’s unclassified systems, with one agreeing across all systems already, and we are optimistic the rest of the companies will get there on classified settings in the near future,” the official said.
The outcome of this dispute could influence how future defense AI contracts are structured. If the Pentagon insists on unrestricted access for lawful uses, technology firms may face increasing pressure to adjust internal safeguards when serving national security clients.
Military AI needs and corporate safeguards
Neither Anthropic nor the Pentagon confirmed whether Claude was used in the Maduro operation.
Advanced AI systems such as Claude are designed to process and synthesize vast amounts of data at speeds beyond human capability. In a high risk overseas mission, that could include analyzing intercepted communications, summarizing intelligence reports, identifying inconsistencies in satellite imagery or cross referencing travel and financial records to verify a target’s location.
Rather than reviewing hundreds of pages of raw intelligence, planners can direct an AI system to surface key insights and highlight potential gaps. These tools can also assist in scenario planning, modeling how weather shifts, route changes or unexpected movements might affect an operation. By integrating logistics data, terrain analysis and known adversary behavior, AI systems can provide commanders with near real time assessments of risks and options.
The broader debate over fully autonomous weapons remains one of the most contentious issues in military AI development. Supporters argue automated systems can respond faster than humans in high speed combat situations. Critics counter that removing human judgment from lethal decisions raises serious legal and accountability concerns if errors occur.
Sources
