U.S. Used Anthropic’s Claude AI in Maduro Capture Raid, Sources Say
Sources tell Axios the U.S. military used Anthropic’s Claude AI during the active operation to capture Nicolás Maduro, and that Claude is currently the only major frontier model available on classified Pentagon networks. Anthropic is negotiating terms with the Pentagon to bar uses such as mass surveillance of Americans and fully autonomous weapons while other firms (OpenAI, Google and xAI) have non‑classified Pentagon deals; Anthropic also has a Palantir partnership that could enable Claude in security products, though it’s unclear if that channel was involved — Fox News described U.S. troops “blasting through steel doors” during the raid.
📌 Key Facts
- Axios reports, citing two sources, that the U.S. military used Anthropic’s Claude AI during the operation to capture Nicolás Maduro; sources say Claude was used during the active raid itself, not only in pre‑mission planning.
- According to Axios, Claude is currently the only major frontier model available on classified Pentagon networks.
- Anthropic is negotiating terms with the Pentagon intended to bar uses such as mass surveillance of Americans and fully autonomous weapons, while the military is reportedly pushing for fewer restrictions.
- Other major model-makers — OpenAI, Google and xAI — already have Pentagon deals for non‑classified use with fewer safeguards and are in talks about access to classified systems.
- Anthropic has an existing partnership with Palantir that allows Claude to be used within Palantir’s security products, though Axios says it is unclear if that channel was involved in the Maduro operation.
📰 Source Timeline (2)
Follow how coverage of this story developed over time
February 13, 2026
10:55 PM
Pentagon used Anthropic's Claude during Maduro raid
New information:
- Axios reports, citing two sources with knowledge, that the U.S. military used Anthropic’s Claude AI system during the operation to capture Nicolás Maduro.
- Claude was used during the active raid itself, not only in pre‑mission planning, and is currently the only major frontier model available on classified Pentagon networks.
- Anthropic is negotiating terms of use with the Pentagon, seeking to bar uses such as mass surveillance of Americans and fully autonomous weapons, even as the military pushes for fewer restrictions.
- Other major model‑makers — OpenAI, Google and xAI — already have Pentagon deals for non‑classified use with fewer safeguards, and are in talks about access on classified systems.
- Anthropic also has an existing partnership with Palantir that allows Claude to be used within Palantir’s security products, though it is unclear if that channel was involved in the Maduro operation.