Pentagon Now Actively Engineering Its Own LLMs After Anthropic Contract Collapse
TechCrunch / Bloomberg·high signal
Pentagon CDAO Cameron Stanley confirmed to Bloomberg that the DoD has begun engineering work on internally developed LLMs and expects them operational 'very soon,' with xAI and OpenAI serving as interim alternatives. Defense Secretary Hegseth's supply-chain-risk designation against Anthropic — normally reserved for foreign adversaries — remains in effect and bars Pentagon contractors from using Anthropic. MIT Technology Review separately reports the Pentagon is also planning to give commercial AI firms access to classified training data.