The Journal. Anthropic’s Pentagon Problems
223 snips
Feb 23, 2026 
Amrith Ramkumar, a WSJ defense and tech reporter, unpacks the clash between Anthropic and the U.S. military. He outlines why Anthropic took a $200M Pentagon contract and Claude’s classified-access perks. He covers limits Anthropic set on surveillance and weapons work, reports about Claude’s alleged role in a Venezuela strike, and the Pentagon’s threat to label the company a supply-chain risk.
AI Snips
Chapters
Transcript
Episode notes
Anthropic’s Safety Brand Enabled Big Pentagon Deal
- Anthropic positioned itself as an AI safety-focused alternative to OpenAI and landed a $200 million Pentagon contract despite that image.
- The deal gave Claude special classified-access via a Palantir partnership, making it uniquely embedded in defense workflows.
Company Guardrails Collide With Military Needs
- Anthropic's terms explicitly ban Claude from being used for domestic surveillance, weapons development, or facilitating violence.
- Those guardrails were written into usage policies and became a direct point of contention with the Defense Department.
Claude Reportedly Used In Venezuela Operation
- The Wall Street Journal reported Claude was used in planning a U.S. strike on Venezuela, prompting internal Anthropic questions about that use.
- An Anthropic employee asked Palantir how Claude was used, triggering alarm and Pentagon awareness.