Overview
- Defense Secretary Pete Hegseth delivered the ultimatum in a Tuesday meeting with CEO Dario Amodei, with reports citing threats to pull a roughly $200 million contract and, unusually, to weigh Defense Production Act powers.
- Anthropic has drawn red lines against domestic mass surveillance and autonomous weapons uses, a stance that contrasts with agreements other vendors such as Google, OpenAI and xAI have made for the Pentagon’s genai.mil network.
- Separately, Anthropic accused China’s DeepSeek, MiniMax and Moonshot AI of large‑scale distillation campaigns against Claude, citing about 24,000 coordinated accounts and roughly 16 million interactions.
- The company attributed query volumes of about 150,000 to DeepSeek, 3.4 million to Moonshot and 13 million to MiniMax, and says it deployed behavioral fingerprinting, output‑degradation for suspected accounts and industry notifications.
- Anthropic also launched enterprise plug‑ins targeting finance, investment banking and HR with partners including FactSet, LSEG, Slack and DocuSign, as markets stayed jumpy after a viral Citrini Research scenario that triggered a broad tech and finance selloff estimated at about $800 billion on its first day.