Pentagon Moves Over Two-Thirds of AI Work Off Anthropic

The U.S. Department of Defense (DoD) says it has shifted more than two-thirds of its daily AI workflows on classified networks off Anthropic to rival vendors. The disclosure came from Emil Michael, the Pentagon's under secretary for research and engineering and CTO, around June 13, 2026, marking a milestone ahead of a full phase-out deadline in September. The transition centers on work running on classified networks, with models from OpenAI and Google stepping in as replacements. The progress has been corroborated by multiple trade reports.