AI on the battlefield: US used Anthropic’s cloud in Maduro operation
Anthropic’s AI model cloud was reportedly used in the US military operation targeting Nicolas Maduro last month. The development places a security-focused commercial AI system at the center of a real-world battlefield mission.
There are moments when technology quietly crosses a line. It appears this is one of them. Anthropic’s artificial intelligence model, Cloud, a system known for drafting large-scale emails, analyzing documents and answering questions, was used in a US military operation aimed at capturing former Venezuelan President Nicolas Maduro, according to people familiar with the matter. In this mission conducted last month, several places in Caracas were bombed and Maduro and his wife were targeted.
Details of how the cloud was used are unclear. Neither operational details nor the exact role played by the AI system have been disclosed. But the mere fact that a commercial AI model has found its way into a live military operation cannot be ignored.
“We cannot comment on whether the cloud, or any other AI model, was used for a specific operation, classified or otherwise,” an Anthropic spokesperson told the WSJ. “Any use of the cloud, whether in the private sector or at the government level, is required to comply with our usage policies, which govern how the cloud can be deployed. We work closely with our partners to ensure compliance.”
The reported deployment of Cloud came through Anthropic’s partnership with Palantir Technologies, whose software platforms are commonly used by the Department of Defense and federal law enforcement agencies. Through this channel, the cloud became part of systems already embedded within the national security framework.
Growing tension between AI security measures and military use
What makes this development particularly impressive is Anthropic’s own rulebook. The company’s usage guidelines prohibit the cloud from being used to promote violence, develop weapons, or conduct surveillance. Yet the operation in question involved the bombing of several locations in Caracas. That tension between written policy and battlefield reality is now at the center of a growing debate.
Anthropic was the first AI model developer whose system was used in classified operations by the Department of Defense. It is possible that other AI tools were used for unclassified tasks in the Venezuela mission. In military environments, such systems can help analyze large amounts of documents, generate reports, or even support autonomous drone systems.
For AI companies competing in a crowded and high-valuation industry, military adoption matters. It signals confidence and technical ability. Additionally, it also brings reputational risk.
Dario Amodei, chief executive of Anthropic, has spoken publicly about the dangers posed by advanced AI systems and called for stronger guardrails and regulation. He has expressed concerns about the use of AI in autonomous lethal operations and domestic surveillance, two areas that have reportedly become sticking points in contract discussions with the Pentagon.
A $200 million contract awarded to Anthropic last summer is now under investigation. Previous reporting has indicated that there are concerns within the company about how the cloud could be used by the military, which has prompted administration officials to consider canceling the agreement.
It appears the disagreement has escalated beyond an operation. This shows deep divisions over how AI should be regulated. The Trump administration has advocated a lighter regulatory approach, while Anthropic has been seen as pushing for stricter safeguards and limits, including on AI chip exports.
At a January event announcing that the Pentagon would be working with XAI, Defense Secretary Pete Hegseth said the agency “will not employ AI models that will not allow you to fight a war,” referencing officials’ discussions with Anthropic.
