Marketplace All-in-One Bytes: Week in Review - Anthropic and the Pentagon face off, OpenAI teams up with consulting firms and Mac Mini moves to the U.S.
Feb 27, 2026
Maria Curie, Axios tech policy reporter who tracks government and industry AI interactions, breaks down a tense Pentagon standoff with Anthropic over access to its Claude model. She also covers OpenAI teaming up with consulting firms to help companies adopt AI coworkers. Plus, Apple’s plan to make the Mac Mini in Texas and what that means for jobs and training.
AI Snips
Chapters
Transcript
Episode notes
Anthropic Is Deeply Embedded In Pentagon Operations
- Anthropic is uniquely entrenched at the Pentagon as the only company providing its model, Claude, for classified use.
- That deep integration makes offboarding difficult and gives both Anthropic and the Pentagon high leverage in negotiations over access and safeguards.
Claude Was Used In The Maduro Raid Without Policy Violations
- Maria Curie cites Claude's use in the Maduro raid as an example of classified deployments that didn't violate Anthropic's policies.
- That suggests sensitive missions can leverage Claude without breaching Anthropic's red lines on weapons or mass surveillance.
All Lawful Purposes Could Enable Mass Surveillance
- The Pentagon's requested "all lawful purposes" access could legally permit continuous AI-driven use of publicly available U.S. data.
- Anthropic objects because legal uses like collecting social media or voter rolls could enable de facto mass surveillance at scale.